Arjun Agrawal · IB DP Economics

Think like an economist.

A free, browser-based learning environment for IB Diploma Programme Economics, built by Arjun Agrawal: IB DP Facilitator in Economics, Business Management and Global Politics, Economics educator, UPSC CSE mentor and resource creator.

Learn the concept. See the model. Build the chain. Stress-test the argument. Make the judgement. Apply it under pressure.

---

What this is

Not a content repository. A system for practising the sequence that IB DP Economics actually assesses:

cause → behaviour → market effect → second-round effect → wider consequence → evaluation → judgement

Everything runs in the browser. There is no account, no tracking, and no AI API. A student can download index.html, double-click it, and work offline. The one piece that runs on a server is optional: the video studio reads the channel's public uploads through a small Vercel function (api/youtube.js), which holds the YouTube key and receives nothing about the reader. Without it, the video studio says so and everything else works exactly as before.

Who it is for

• Students taking IB DP Economics at SL or HL, at any point from first lesson to final exam.
• Teachers who want accurate diagrams, classroom-projectable models, misconception tools, question banks and assessment reasoning they can use directly in a lesson.
• Anyone curious about how the economy works: school and university students, parents, professionals. Economics, Everywhere is written for them: everyday questions answered from the mechanism up, without assuming the IB course.

Economics is not only a subject you study. It is a way of understanding the world.

What it contains

Section	What it does
Course	The reference layer: the course at a glance, the six real-world issues, the nine key concepts, a complete curriculum explorer reading every subtopic three ways, an SL/HL difference explorer, a visual knowledge map, a 250-term glossary (Glossary 2.0, under Course › Dictionary) and a command term lab covering all 33 terms in the guide's glossary.
Learn	Every concept in nine stages: the idea, intuition, the model, the diagram, the chain, the application, the stress test, the IB connection, retrieval. Plus a curriculum map and a misconception lab.
Think	The Economic Chain: build a mechanism link by link against a live shock, then survive a conditionality test and choose between four judgements. Plus an evaluation stress test across fourteen dimensions, a Think › Dimension library setting out those dimensions as Evaluation is conditionality, a concept-connection exercise, and a conditions library of the twelve conditions that carry almost every "it depends" in the course, with an assumptions lab and a "when does the model break?" lab beside it. Last, the economist's toolkit: eighteen lenses (incentives to institutions), each with its questions and links into the platform's own material.
Lab	Computed market and AD-AS models you can move, a transparent policy simulator, a 35-plate diagram atlas (each plate with a reading guide: what changes, why, what the diagram does not show and the common mistake), a counterfactual "What if?" lab, a policy Decision Room, a "build it yourself" diagram builder, a 20-policy toolkit with a comparison matrix and a two-policy comparator, a Recommend-a-policy lab for HL Paper 3 part (b), an 11-model library stating assumptions before diagrams, a stakeholder / time-horizon / context lab, a predict-before-reveal mode for the diagrams, a "check my diagram" error detector, and three live integrity dashboards: model, content and links.
Calculate	The calculation centre: 36 calculations with live working, units, interpretation, common mistakes, linked diagrams and generated practice that hides the solution until you attempt it: plus a printable calculation sheet for the student's own level.
Real World	The largest section on the platform. 203 real-world cases remapped to the 2022 guide, each carrying its subtopic, level, papers, key concepts, diagram and evaluation, searchable and filterable by topic, concept, assessment, place, theme, level and case type; twelve of them worked to a full teaching standard. Around them: a twelve-zone case reader, a case-connection graph, Find me an example, Build an evaluation, Compare two cases, Use this case, three transfer challenges, My casebook, a syllabus atlas, a concept index, an assessment matrix and a sources-and-method page. The earlier surfaces are all still here: six recurring economic events, the eight worked cases split into verified fact, interpretation and inference, a data sources library, and an "Economics now" page that states plainly that this platform holds no live data.

Design and discovery

The platform reads as an economics publication rather than a long page of cards.

• An opener that is about the page. Every page opens with a figure drawn for what it covers: comparative advantage on the comparative advantage page, a Lorenz curve for inequality, markbands in the exam room, and a case's own diagram on each Real World case. The 28 figures are built from stated functions, so their intersections, wedges and shaded areas are where the economics puts them, and the self-test recomputes them.

• A register for each section. Journal paper for the course, Learn, the IA, the EE Studio, TOK and Educators; a newsroom for Real World; graph paper for the labs; ruled paper for the exam room; a field of nodes for mindmaps; a magazine for Economics, Everywhere. These are one system with one palette (obsidian, burgundy, chalk, and brass for rules and ornaments). Semantic colour stays inside diagrams.

• The brand, in vector. The logo, reversed logo, monogram, wordmark, favicon and app icons are SVG masters in assets/brand/, crisp at any size.

  • Monogram: traced from the supplied artwork and fitted to straight lines and curves.
  • Wordmark: set in Cinzel (SIL Open Font License), converted to outlines on the original letter positions.
  • In the page: the embedded vectors replaced the PNGs, and the PWA icons are rendered from the masters.

• The opening. The brand line, Think like an economist., is the typographic signature.

  • Behind the logo, seven ideas (price, quantity, incentives; markets, trade, money, policy) are linked by lines that fade in once.
  • It is static under reduced motion and left out on phones.

• The home page. It opens on a cover and a working model.

  • The cover. Think like an economist. leads, with a one-line statement, five verbs (Explore, Learn, Apply, Question, Master) and three actions.
  • A market you can move. A slider shifts demand; the plate redraws, and one sentence explains the shortage or surplus that moves price. Six concept chips show their definition on hover or focus.
  • The author. A tall editorial panel in its own column, with an author line taken only from the supplied profile. Nothing is drawn behind the photograph at any width.
  • The platform in numbers. Cases, mindmaps, plates, calculations and Exam DNA questions, set in the data face along the foot. They are counted from the data, and each opens what it counts.
  • Eleven chapters: Think, The course, Learn, See the world, Master the models, Exam, Research, Explore, Watch, Teach and About.

Economics, Everywhere

A section for readers who are not taking the IB course, and for IB students who want the economics behind it. It lives at #/everywhere and is linked from the bar, the drawer, the More menu and the home page.

Every big question is built the same way. The question → the short answer → the economic idea → Why does this happen?, the mechanism revealed one step at a time → a real example, labelled → Change one thing, an interactive lab → What would you expect?, a prediction made before the reasoning is shown → Go deeper, at beginner, student, IB or advanced level → the confusion to avoid → links to the platform's concepts, cases, mindmaps, diagrams and calculations → a pathway onward → sources.

Every claim carries its layer. A fact names its source, drawn from one register of checked URLs (central banks, statistical agencies, the IMF, World Bank, OECD, WTO, ILO, UNDP, the Nobel Prize site, and a small number of papers and reputable reports). An illustrative example uses invented numbers and says so. A stylised model is labelled on every lab and states its assumptions; none is a forecast. Nothing in the section is live data: Where the numbers live says which institution publishes each current figure and how to read it, and holds no figures itself.

The labs are declared models checked against independent solutions. Price, inflation and purchasing power, tax incidence, the minimum wage (competitive and monopsony), the exchange rate, interest rates and loan repayments, externalities, comparative advantage, a repeated pricing game, a public-goods game, who gets the growth, the power of growth, a membership calculator, an AD-AS explorer and network effects. The self-tests check, among other things, that tax incidence follows PES / (PES + |PED|), that the exchange-rate lab switches exactly at the Marshall-Lerner condition, that monopsony employment traces an inverted U, and that both countries gain from trade exactly when the terms of trade lie between the two opportunity costs.

It remembers, on this device only. Saved pieces join the platform's single saved list; recently opened pieces appear under Continue exploring and in the home page's continue-learning record; the chosen explanation level is remembered. Nothing is transmitted.

Economics EE Studio

A workspace for the Economics extended essay: Research Economics. Build an argument. Think like an economist. It works as diagnose → question → refine → test again. It never writes, approves or grades anything, and everything a student enters stays in their own browser.

Source discipline. Every statement about the essay carries one of five labels: Official IB requirement (read first-hand in the Economics guide or its teacher support material, with the page); Official IB requirement · reported (from the Extended essay guide for first assessment 2027, which was not supplied to this project, via the author's Extended Essay Navigator: to be confirmed in the guide); Professional interpretation; Pedagogical suggestion; Teacher-created tool. Where the sources disagree, both are shown: the Economics guide (2022, p. 8) says the issue "must have taken place up to five years prior to the beginning of the research process", while the Navigator reports no recency rule in the 2027 Extended essay guide. The studio says the current Extended essay guide governs and sends the student to check it with their supervisor.

Arjun Agrawal · Video studio and the YouTube feed

The video studio (#/arjun) lists the channel's videos automatically. No video is written into the page: a new upload appears on its own.

How it works. The page calls /api/youtube, a Vercel serverless function in api/youtube.js. With a key it reads the channel's uploads playlist, durations (to identify Shorts) and playlists through the YouTube Data API v3, resolving the channel from its handle; without a key but with a channel ID it reads the channel's public RSS feed (the latest 15 uploads, no playlists or durations). The answer is cached in the function's memory for 30 minutes and at Vercel's edge for an hour, so YouTube is not called on every page view; if YouTube fails, the last good answer is served and marked as stale. Every call to YouTube has a timeout and one retry. Categories come from, in order: manual overrides, playlist titles, then words in each video's title and description (lib/youtube-categories.js; add a category by adding one entry there).

Setting it up (Vercel). In the Vercel dashboard open the project → Settings → Environment Variables, add the following, then redeploy:

Variable	Required	What it is
YOUTUBE_API_KEY	Recommended	A YouTube Data API v3 key. Create it in the Google Cloud console (APIs & Services → Credentials), enable "YouTube Data API v3" for the project, and restrict the key to that API. It is read only by the server function and never sent to the browser.
YOUTUBE_CHANNEL_HANDLE	No	Defaults to @arjunagrawal5724.
YOUTUBE_CHANNEL_ID	No	The channel's UC… ID. With a key, it skips the handle lookup. Without a key, it switches on the keyless RSS fallback.
YOUTUBE_TIMEOUT_MS	No	Timeout for each call to YouTube; default 8000.

With neither a key nor a channel ID, /api/youtube answers "not configured", and the studio shows a short note with a link to the channel instead of breaking. Quota use is small: one refresh costs about four units plus one per playlist, against the free daily allowance of 10,000 units.

The channel ID was not verified when this was built: YouTube could not be reached from the build environment, so the function resolves the channel from the handle at run time. Once deployed, open /api/youtube and check that channel.title is correct; setting YOUTUBE_CHANNEL_ID then pins it.

Privacy. The page requests only /api/youtube, without cookies or any information about the reader, and never during the self-test. Videos open on YouTube, under YouTube's own terms.

Educational philosophy

A definition recalled without its mechanism is inert. A diagram drawn without an argument attached to it earns almost nothing. So the platform puts the mechanism first, and everything else attaches to it:

Understand → Visualise → Connect → Analyse → Stress-test → Judge → Apply

Around that spine sits the structure the guide itself uses: six real-world issues, read through nine key concepts, across four units. Every subtopic in the curriculum explorer answers the same three questions: what do I need to know, what must I be able to do, and how could this be assessed.

That framework is a teacher's framing, not IB terminology, and the platform labels it as such.

Honesty rules the platform holds itself to

• Five claim layers, never blurred: official IB information · established economic theory · Arjun Agrawal teaching frame · teacher-created practice · constructed teaching data. Assessment claims carry a source, a version and a verification date. Nine of them are also checked against the IB's own pages, with a link to each, under Exam → IB reference layer → Live verification. One of the nine could not be established and is shown as not verified rather than quietly dropped.
• No fabricated data. No figure is asserted in the Real World section; every story routes its numbers to named primary institutions. All lab and drill figures are labelled as constructed teaching data.
• The diagnostic is rules-based, teacher-designed, and built on the published assessment framework. It is not an official IB mark and the platform says so wherever it appears.
• No fabricated social proof. No testimonials, ratings, student counts, success rates or grade guarantees anywhere.
• No invented credentials. The About page contains only details supplied by Arjun Agrawal.
• No universal definitions where none exist. Definitions of recession, for instance, are presented as varying between countries and institutions rather than as a single settled rule.
• Gaps are declared. The diagram atlas lists the plates the guide names that this build does not yet draw, and the quality report lists what the self-tests do not cover, rather than implying coverage that does not exist.
• Nothing pretends to be live. There is no data feed. The "Economics now" page says so, and routes every current figure to the body that publishes it.

Technology

Where the source lives. index.html is still the shipped file, but it is now built:

• src/modules/*.js and src/styles/*.css are spliced into it by python3 tools/build/splice.py. The splice is idempotent, and its output was checked byte-identical before the move.
• Content data is generated into assets/data/ by tools/build/build_ideas.py and tools/build/build_data.py.
• Changes to the older base code are made by recorded, re-runnable scripts in tools/migrations/.

docs/architecture-decision.md gives the reasoning, the measured baseline and the staged plan.

• A single self-contained index.html. No framework, no bundler, no dependencies to install.
• One file, deliberately, with one exception. Everything the platform does is in index.html, because that is what makes it work from a memory stick, from an email attachment and from a file:// path with no server. Parse cost was measured rather than assumed, over three cold loads of the packaged file: first contentful paint 280 to 288 ms, DOMContentLoaded 1.66 to 1.70 s, 964 DOM nodes after the home page renders, a 20 MB heap, and the slowest render in the build is the diagram atlas, now 35 plates. Sixty navigations across the routes (34 sections and 194 tabs) leave the heap unchanged. The parse figure grew with the bundled copies of the two corpora, which is the cost of a file that still works when it arrives on its own. Splitting the remaining datasets would have traded that for a tree of files that breaks the moment one of them is lost. The one exception is the pair of reference corpora, assets/data/exam-dna.js and assets/data/real-world.js, which are loaded with a plain <script src> because they are reference data rather than behaviour and because each must be auditable on its own. Neither is required, and neither can leave a gap. Both also travel inside index.html, as JSON held in function bodies the engine skips until they are called. If the copy in assets/data loads, the bundled copy is never read; if it does not, the platform adopts the bundled copy and nothing on the screen changes. Served beside its assets folder, the build runs the full 2,076 checks with nothing failing. Opened on its own, without assets/, it still works but the checks on the lazily loaded files fail, by design.
• Diagram Engine 2.0. Every quantitative diagram is generated from a declared model: equations, parameters, an equilibrium solver, then curves plotted from those same equations. No intersection is positioned by hand. After rendering, the engine measures the distance in rendered pixels between every labelled point and the curves it is supposed to lie on; a point that misses fails validation and the diagram is reported as failed rather than displayed as correct. 28 model-driven diagrams, 28 validated, 0 failed, largest miss 0.03 px against a 0.75 px tolerance. The seven remaining plates are flow diagrams and schematics with no equilibrium to solve; they are declared as such rather than counted as validated.
• The equations, parameters, solved equilibria and per-point validation for every diagram are visible to teachers at Lab → Model integrity.
• The typefaces (Bodoni Moda, Source Serif 4, Manrope, JetBrains Mono; SIL Open Font License) are served from assets/fonts/, so no font request leaves the site. Nothing else is fetched.

Mindmaps

The knowledge map shows how syllabus topics connect to one another. A mindmap shows something different: how an economist moves through one of them, from the concept to the mechanism to the model to the judgement, and where each step can be tested. Both are kept, and they link to each other.

Forty-two maps in three classes:

• Ten core maps, one per strand of the course: foundations, markets, elasticity, government intervention, market failure, market structure, economic activity, policy, the global economy and development. These are the original ten, retained unchanged in identity.
• Twenty-two topic maps, one per syllabus subtopic that carries enough structure to deserve its own, from demand and supply through to behavioural economics.
• Ten synthesis maps, which exist because the examined skill is connection rather than recall: elasticity and taxation, elasticity and intervention, externalities and sustainability, the interest rate as one instrument across three units, exchange rates and trade, growth and development, growth and inequality, growth and sustainability, efficiency against equity, and scarcity as the thread through the whole course.

Together they cover all thirty-one subtopics in the guide, which the coverage dashboard measures rather than asserts.

Each map is a tree of nodes plus cross-links that cut across branches, which is what lets the elasticity map reach tax incidence without redrawing elasticity inside the tax branch. Eighteen node types carry a consistent visual language across every map, with a distinct shape and colour each: concept, determinant, mechanism, model, diagram, calculation, application, evaluation, stakeholder, judgement, TOK, exam, assumption, formula, misconception, evidence, real-world case and policy.

The layout is computed, not positioned by hand. Branches are dealt to two sides of the central concept and balanced by weight, so every map comes out landscape rather than as a tall column, which is the shape a projector wants. A single renderer draws all forty-two, so adding a forty-third means adding data and changing no code. Seven hundred and sixty-two nodes in total.

The thirteen modes:

Mode	What it asks
Explore	Open any node. Detail, syllabus subtopic, diagram, calculation and dictionary entry appear beside the map.
Follow the mechanism	One causal pathway highlighted, the rest faded, read end to end.
Test yourself	Four nodes blanked; the distractors are real nodes from the same map, so the shape of the topic will not carry you.
Build the map	A pathway shuffled and reordered from memory.
Explain it	Say a branch in your own words against a sixty-second timer, then self-check against four criteria.
Repair it	A claim about the topic that is nearly right. Find the fault, correct it, and send it to the mistake book.
Connect	Build the chain that links this topic to another part of the course, in the order the argument requires.
What changes the result?	Move the condition the conclusion rests on and read what happens to the judgement. The point of the mode is that a conclusion which survives every condition was never an evaluation.
Exam	The questions the map's branches are actually assessed by, with the command term crosswalk.
Real world	One sourced application with its interpretation, model, limitation and publishing institution.
TOK lens	The knowledge question the topic raises, marked optional.
Classroom	One branch at a time, sized for a projector, with a timer and an exit question.
Revision	The map reduced to what has to be recoverable under examination conditions, ordered by what the progress record says is weakest.
List	The whole map as text. Not an afterthought: it is the default on phones and what screen readers and printouts use.

Beneath every mode of every map sits the evaluation engine: the conditions under which that topic's conclusion would change, each one to be written out with a direction, because a condition named without a direction earns nothing.

Mastery is earned, not opened. Five strands per map (explored, mechanism followed, retrieval, reconstructed, explained), and opening a map can move only the first. A self-test asserts that a map with every node opened and nothing else done still scores below secure. Everything feeds the existing skills model and revision queue rather than a parallel one: a branch marked difficult goes to the mistake book, a branch set aside for later goes to the revision queue, and the daily challenge draws from the maps on a rotation.

Validation. Every node reference is resolved against the thing it names. A node pointing at a syllabus code, a diagram or a calculation that does not exist fails the build, and so does an orphan node, a cycle, a cross-link leaving its own map, a pathway naming a node that is not there, an objective using a vague verb, a real-world section asserting a figure of its own, or a map laying out taller than it is wide. That suite found several defects during development, including twelve nodes mapped to the wrong syllabus subtopic and three cross-links pointing out of their own map.

Every mindmap that can carry a case does. Thirty-seven of the forty-two show one to three real-world cases from the collection at their own subtopics, with an Explore case button into the full record. The five that cannot say so rather than borrowing a case from somewhere else.

The eighteenth node type is in use. "Real-world case" asks where something has actually happened, which is a different question from "Application", where the theory is used. Rather than assert a new fact inside a map, each case node names a case the platform already holds in verified form and hands the student to it, so the claim and its sourcing stay in one place.

Stored progress was carried across, not discarded. These were called brain maps in the previous release. The storage key moved with the name, and the migration copies the old record into the new one rather than starting anyone from zero. A test seeds a profile in the old shape, runs the migration and asserts that every strand, every score and every unrelated key survives.

Video learning

Thirty-four curated external videos across nineteen of the thirty-one subtopics, from nine channels. The layer exists because watching is the cheapest form of study and the easiest to mistake for progress, so nothing here is marked as learned because it was played.

Every entry states the subtopic it is filed against, whether the content is SL or HL, what it is for (concept, diagram, calculation or revision), and why it was chosen rather than something else. Beside each one sit three checkpoint questions, written against the syllabus content of that subtopic rather than against the wording of the video, and the panel says so: a question written from a transcript tests whether you were paying attention to a presenter, and a question written from the guide tests whether you can now do the thing the subtopic asks for.

Nothing loads until it is asked for. The grid fetches still images only. The player element does not exist anywhere in the document until the student presses Load the player, and it is then created against YouTube's privacy-enhanced domain and never autoplays. A test asserts that both the closed library and the opened viewer contain no player element at all.

No broken image, ever. Each still has two hosts and then a drawn placeholder generated inside the page, which needs no network. A watchdog replaces any still that was requested and has not arrived after five seconds, which matters because many schools block the video hosts outright.

Verification is a build-time fact, stated with its date. Every identifier was confirmed against YouTube's public oEmbed endpoint, which returns a title and a channel for a video that exists and is publicly available and an error for one that is private, deleted or restricted. The titles and channels recorded are the ones that endpoint returned, not the ones a search result displayed. Running times are not printed, because that endpoint does not return one and a figure that has not been checked is not worth printing.

Videos are external resources selected for conceptual support. They are not official IB materials unless explicitly identified as such, and none held here is so identified.

The eleven tools

Tool	What it does
Diagram repair shop	Twelve cases, each a candidate's diagram described exactly as it was drawn. Find the fault, name the correction, then compare with the atlas version. The corrected diagram is the atlas diagram, so there is only one drawing of each model in the platform.
Data detective	Eight tables, each with three claims: one the data support, one they contradict and one they cannot settle. Naming what a table cannot show is what the evaluation band rewards.
Policy memo lab	Six briefs, each written up in six sections: situation, objective, instrument, mechanism, risks, recommendation. A memo will not let you recommend before you have explained.
Economic debate room	Ten motions. Both benches must be argued before the judgement opens, because an evaluation that has only ever argued one side reads as a list of disadvantages.
Market scenarios	Eight named markets loaded into the market simulator that already exists, with the parameters that make each one behave as it does. Not a second simulator.
Comparative advantage game	Two countries, two goods, four output figures. Everything else (the opportunity costs, which country has the comparative advantage in which good, the range of terms of trade at which both gain) is computed from them.
CPI basket builder	Ten weighted categories. Change a weight or a price and watch which categories actually move the headline, and how a weighted index differs from an average.
Lorenz curve builder	Five quintile shares in, a curve and a Gini coefficient out, computed by the trapezium rule with the working shown. The grouped-data ceiling of 0.8 is stated rather than hidden.
Balance of payments ledger	Fifteen transactions posted to the sub-accounts the guide names, with the right sign. Posted correctly the whole ledger sums to zero, which is what the identity means.
Exchange rate lab	Seven named causes shift demand for or supply of the currency. The Marshall-Lerner condition falls out of the two elasticities entered rather than being asserted beside them.
Multiplier lab	Set the withdrawals, read the coefficient, and follow an injection round by round until it converges on the closed-form total.

Timed sessions

Revision fails in two directions. It fails when a student has ten minutes and opens nothing because nothing fits ten minutes, and it fails when a student intends twenty minutes and spends ninety re-reading. Naming the length first, and building the work to fit it, is what stops both.

Five lengths: five, fifteen, thirty, sixty and ninety minutes. Five minutes is three retrieval items, enough to keep a topic warm and not enough to learn one; ninety minutes finishes with an original examination task written under the clock. Each session is assembled when it starts, from the subtopics the progress record shows as weak and the mistakes still open in the mistake book, so two sessions of the same length are not the same session. The suggested minutes are scaled so that the steps add up to the length chosen. Self-assessment is recorded at a lower weight than a marked attempt, and the panel says so, because it is the student's reading of their own work.

Exam DNA

A released examination paper is somebody else's copyright. The pattern inside it is not.

Exam DNA reads 212 released questions and their 990 parts, across twelve sessions of Papers 1, 2 and 3 at both levels, and records for each part only its syllabus subtopic, its command term, its mark tariff and what the task demands: a diagram, a calculation, use of the data, evaluation, a real-world example, a policy. It holds no question text, no stimulus, no markscheme and no paraphrase of any of them. The corpus was machine-checked against the source papers with all whitespace removed, and the longest run of consecutive words shared with any paper is zero.

What that buys a student is the shape of the paper rather than its contents:

• The explorer filters every recorded part by paper, level, session, subtopic, command term and demand, and reports how the marks actually fall.
• Eleven archetypes name the recurring task shapes, from the ten-mark mechanism to the Paper 3 recommendation, each with the demands it carries and the trap it sets.
• The mark lens shows what a two, four, ten or fifteen-mark tariff has historically asked for, so a student stops writing the same paragraph at every tariff.
• The heatmap puts subtopics against demands, which is where a revision plan should start.
• The morpher takes a recorded shape and builds an original task from it, written here. It is not a past question and the panel says so.
• The museum shows how the assessment has moved across the sessions analysed.
• The source register names all five papers analysed, what was taken from each and what was not, and links to the official IB Questionbank and to resources.ibo.org rather than hosting anything.

No prediction is made anywhere. The corpus records what was asked, not what will be. Every surface carries the same notice, and the analysis date is the date the papers were read.

Three confidence levels are recorded rather than hidden: 803 parts classified directly from the task, 122 from the stimulus, 62 inherited from the question stem and 3 left unclassified. A classifier that cannot tell says so.

Exam practice

Original examination material, written for this platform and labelled as teacher-created practice. Nothing here is an IB question and nothing here is described as one.

• Paper 1: eight extended-response tasks, each with the command term, the mark split, the diagram the question expects and the evaluation the top band asks for.
• Paper 2: four data sets with their own stimulus, twenty-eight parts between them, built so that the text carries information the questions actually need.
• Paper 3: four HL cases with twelve calculations, each answer recomputed from the table beside it rather than asserted.
• The simulator: five runs, from a ten-minute retrieval drill to a full mock, with a clock that counts down and then counts up, autosave, and a review that compares what was written against the demands the task carried.

All figures are constructed teaching data and are labelled as such.

Real-world economics

Economics becomes easier to understand when you can see it happening.

The largest section on the platform, and the one the rest now connects through. 203 cases, remapped to the 2022 guide, held in one file and referenced from everywhere else. A case exists once. Nothing copies it.

What a case carries. Its 2022 subtopic code, its unit, its level status (core, carrying HL depth, or HL only), the papers it serves, its key concepts with the primary lens first, the economic issue, the mechanism, the diagram it actually supports, the evaluation opening, its Paper 1, Paper 2 and possible internal-assessment use, a discussion question, an examiner-style application note, its region, its themes, its case type, and the first-edition section it came from.

Six doors into the same bank. By topic, by key concept, by assessment, by country, by theme, by level. All six narrow one collection rather than opening six of them.

Search that says why it matched. Searching price ceiling, PED, India or equity returns cases with the reason each one matched printed under it: the title, the country, the syllabus code, the concept, the diagram or the evaluation line. The same cases are in the global command-palette search, filed under Real world, which now carries close to 2,900 rows across more than sixty kinds of content.

The twelve-zone case reader. Context · the economics · model · diagram · stakeholders · evidence · limitations · evaluation · key concept · exam use · IA use · TOK. One zone at a time, with a before-and-after panel and an eight-step policy chain in the economics zone, a who-gains-who-loses prompt in the stakeholders zone that refuses to pretend the answer is universal, and a data-freshness banner on every case.

Twelve deep teaching cards carry the full field set: context, theory, stakeholder table, arguments for and against, the balanced verdict, common student errors, per-paper deployment, a classroom kit with hook, starter, core activity and exit ticket, the misconception to intercept, an extension task, the examiner's lens, a TOK link, and a forty-word student revision rail with a memory trigger and an exam tip.

A connection graph on every case, resolving to records that already exist: the subtopic, the mindmaps covering it, the verified diagrams, the calculations, the curated videos, the exam archetypes, the dictionary terms, the TOK knowledge question, and four cases worth comparing it with.

Five case tools.

Tool	What it does
Find me an example	Choose topic, concept, paper, place and level. Returns best-fit cases with the reason each fits, banded as a strong match, a useful match or worth a look. No case is ever called the best.
Build an evaluation	Seven moves against a chosen case: condition, time, magnitude, stakeholder, evidence, limitation, judgement. Printable. Nothing is marked.
Compare two cases	Two cases side by side, with what is the same, what differs, and the lesson. Nine ready-made comparison sets are built in, from the hyperinflation trio to the carbon-pricing family.
Use this case	An original examination-style task and five slots: claim, mechanism, evidence, evaluation, judgement. The platform does not write the answer.
Case challenges	Case to diagram, case to mindmap, case to command term. Three options, one correct, and then the sentence that has to follow it.

My casebook. Saved cases, using the saved-items store that already exists rather than a second one, with a coverage read-out: which units are thin, which concepts are missing, and why that matters for a paper that offers a choice of three questions and for an internal assessment that requires three different units and three different concepts.

A teacher view of every case, with a learning objective, best classroom use, a starter, a core activity, the diagram, discussion, evaluation, the misconception to intercept, an extension, an exit ticket and a time estimate, and a printable lesson card.

Where the numbers come from. No figure is presented as current. Every case states how current it is (recent episode, settled episode, historical data), and the evidence layer names the national statistics office, the central bank, the ministry or regulator, the international institution and, last, the reputable secondary source, with what each one settles. The standing note from the source collection is kept: the durable content of a case is its mechanism, its diagram and its evaluation, not its headline number.

What it does not do. It makes no prediction. It never says a case is worth a number of marks. It does not supply an internal-assessment article, and it says so on every IA surface: the article has to be current, published and found by the student. Development lives in Unit 4, where the 2022 guide puts it, and Development is a theme inside that unit rather than a fifth unit of its own.

Declared gaps. 24 of the 31 subtopics hold at least one case. The seven that do not are named on the atlas, in the coverage dashboard and on the sources page, alongside the four commissioning priorities the source collection itself identified: public goods, the HL equity subtopic, inequality and poverty measurement, and Unit 1 methodology. A gap is stated rather than padded.

The learning chain

Every major concept now connects through a visible chain, and each link opens a record that exists once:

topic → concept → mindmap → diagram → calculation → misconception → real-world example → video → TOK → Exam DNA → original practice → revision → progress

• 37 of the 42 mindmaps carry one to three real-world cases; the five that do not say so rather than borrowing one.
• Every lesson carries its real-world cases under The real world, and an inquiry.
• The home page features one case, chosen from the subtopics the student's own record shows them working on, with a concept of the week and an original question of the day that rotates across recall, calculation, diagram, evaluation, real world and TOK.
• The coverage dashboard now counts real-world cases, curated videos and analysed question parts per subtopic, so a gap in any of them is visible in one table.

Sticky navigation

The top bar, the section tabs, the lesson rails and the events chapter rail stay in view while the page scrolls. They were always styled to do so, but body carried overflow-x:hidden, which (with html clipping sideways overflow) made body a scroll container that never scrolls, so nothing ever stuck. body now uses overflow-x:clip, which still prevents sideways scrolling without creating a scroll container. The widths audit also now fails on content cut off at the right edge, which that clip would otherwise hide.

Navigation

The primary bar keeps the eight places most visits start from: Course · Learn · Lab · Real World · Practise · Exam · Mindmaps · Resources, with More and Tutorials closing it. Resources opens the Resource hub (#/resources). Economics, Everywhere and the Learning Studio are in the More menu, the drawer, the footer and search, and the home page opens a door to each.

The bar is measured after every paint, and again once the typefaces arrive. An item is withdrawn only when it would otherwise meet the brand, in a declared order; Resources is the last to go, then Real World and Mindmaps. At 1440px and wider all eight fit. To give the bar that room, the tools beside it fold:

• The timer is a single stopwatch control while it is idle. It opens the paper presets, and the readout and pause control return once it runs.
• The profile shows its initial, and keeps its name for assistive technology.
• Below 1180px, where the bar gives way to the drawer, a Resources button stands in the header between the brand and the menu. Below 360px it gives the wordmark its room, and Resources leads the drawer instead.

Everything else is one step away:

• The More menu groups every section by what a visitor came to do: Learn and explore (Resources first, then the course, Think, Why did this happen?, economists and ideas, economic data, the labs and the video studio), Exam and research, Your work, Teach and About. Any section not named there is appended under More, so nothing can become unreachable.
• On a phone, the drawer carries the same sections in its own groups.
• Search (/ or Ctrl/Cmd + K) finds lessons, terms, economists, causal questions, data pages, cases and tools by name. A calculation's result, like its address (#/calculate/0/ped), opens that calculation; simulator and worked-case results open their tabs.
• A built-in calculator opens from the header's Open calculator button and takes keyboard input.

Learning paths and saved items

One save control and one difficulty flag, defined once and used by the mindmaps, the video library, the topic dossiers and the tools. Flagging something moves its subtopic to the front of the spaced revision queue: the flag is about the economics, not about the resource.

Seven paths, each of which opens material that already exists, in an order chosen because the concepts depend on one another: microeconomics from the beginning, macroeconomics from the beginning, the global economy, Paper 1 in two weeks, Paper 3 at HL, the internal assessment start to finish, and the week before the examination. A path is a suggested order rather than a requirement, and marking a step done records that you opened it and nothing more.

The assignment builder

A teacher assembles mindmaps, subtopics, videos, tools, calculations, diagrams and learning paths into one sheet, prints a student version and a teacher version, or copies a link that carries the whole assignment inside itself.

There is no server, no account and no class list. The link contains the chosen items and nothing else: no names, no progress, nothing about any student. A student who opens it sees a panel offering to open the assignment, and nothing about what they then do is reported to anyone.

The opening sequence

Exactly five seconds, on one clock. The sequence runs:

Time	Stage
0–0.9 s	an obsidian field with a faint grid; the axes rule themselves in from their origins
0.7 s	the monogram, with a data series appearing beneath it (0.9–1.9 s)
1.4 s	the name and "Knowledge. Insight. Impact.", assembled from the brand vectors rather than set again in type
2.2 s	"IB DP Economics", with supply and demand and a frontier (micro) brought into focus behind it
3.1 s	"Think like an economist.", the peak, as a growth path with its cycle (macro) and a faint network of trade (global) take over the background
4.0 s	"Learn. Think. Calculate."
4.7–5.0 s	the exit fade into the page

One start timestamp. The start is the first frame the reader sees. It is taken from requestAnimationFrame and re-anchored to the browser's reported first paint during a still passage of the sequence. Every stage is a Web Animation scheduled from that timestamp, not a chain of timers.

Nothing can lengthen it. The stages animate opacity and transform only, so they run on the compositor while the start-up self-test occupies the main thread. The exit fade reaches full transparency at exactly start + 5000 ms, and the overlay is removed on the next frame.

Fallbacks rather than delays.

• The font stylesheet no longer blocks the first frame; a late font is swapped in, not waited for.
• On a device slow enough that the platform is not ready at 4.7 s, the final frame holds rather than uncovering an unfinished page, and the timing record says so.

When it plays: the opening state model. The rule depends on the address, not on the cache or the kind of load:

How the reader arrives	What they see
The home page (/, #/home), first visit or any later one	the full five seconds
The home page by reload, hard reload, cache disabled, a bookmark or a typed address	the full five seconds
The home page by the browser's Back or Forward from another site	the full five seconds
The home page from a page restored by the back/forward cache	the page as it was left (it never unloaded, so there is nothing to open)
A deep link to any other page (#/events/archive/india-1991)	the identity still, gone as soon as the page is ready
Moving inside the platform, including Home and Back between its own pages	no opening at all

The page records the navigation type, the route and the result in window.__INTRO. There is no timer that waits for the opening: the page is built behind it, so it is ready when the overlay leaves.

Skip and reduced motion.

• Skip intro fades out over about 170 ms and leaves through the same exit.
• Reduced motion shows the same hierarchy at once and leaves after 1.5 s.

tests/intro.mjs measures the timing at 360, 390, 768, 1024, 1440 and 1920 px, from the record the page keeps in window.__INTRO, and every row of the table above: a fresh profile, reload, hard reload with the cache disabled, a direct address, Back and Forward from another site, a revisit controlled by the service worker, and a deep link.

Installable, and honest about offline

The whole platform is one self-contained file, so once it has loaded on a device everything it does itself works without a connection: every mindmap, every diagram, every calculation, every dossier, every tool, every question and every printable, and the progress record, which lives in that browser and nowhere else.

What does not work offline is everything that belongs to somebody else: the videos and their still images, which are served by YouTube, and the resource library, which is held in Google Drive. The service worker leaves other origins entirely alone and never answers a request to them from a stale copy, and the offline panel on the About page says exactly this. Installing adds an icon and opens the platform in its own window. It creates no account, signs you in to nothing and sends nothing anywhere.

The landing experience, the Learning Studio, Learning design and Contact

The home page leads with the platform's signature, "Think like an economist.":

• The idea: "Economics is not just a subject to study. It is a way to understand the world."
• What the platform is: "An interactive IB DP Economics learning ecosystem connecting economic concepts, visual understanding, real-world evidence and exam thinking."
• Two actions:
  • Explore Economics moves down to Explore Economics Resources.
  • Explore Resources opens the Resource hub (#/resources).

The author signature sits level with the foot of the figures. It shows:

• the original photograph, cropped to 4:5 (assets/brand/arjun-agrawal-portrait-4x5.webp, made from assets/brand/arjun-agrawal-portrait.jpg, which is unchanged);
• the name, and "IB DP Economics Educator and Resource Designer";
• links to About and to the Learning Studio.

The figures beside the words are three, numbered:

1. The live market model. It is drawn taller, so it fills its figure.
2. The corrected elasticity diagram, with the same axes and the same scale.
3. The revenue diagram from the same explainer.

The two excerpts share the model's height. Between 900 and 1100px only the model is shown beside the words. On a phone the order is statement, idea, actions, model, then the signature. The page then reads as follows.

• Explore Economics Resources, directly after the cover. It has eight doors, each a real page with a small figure and a count taken from the platform's data:

  • topic lessons;
  • mindmaps and visual explanations;
  • Economics diagrams;
  • real-world examples;
  • calculations;
  • exam practice and revision;
  • Economics in 60 Seconds;
  • Extended Essay support.

  The Resource hub's library follows, listing its own eleven categories, then Random Economics, Find a topic and the economist's toolkit. A single line points to the Learning Studio, which stays separate: the hub is the free library, and the Studio shows selected previews of original work. This band takes the place of the earlier six pathways and gateway, which led to the same pages.

• Inside the Economics Learning Studio. An exhibition of five pieces in five forms:

  • a featured spread (exchange rates);
  • a visual explainer (PED);
  • an editorial guide shown as pages (the examiner's guide);
  • an application shown by its title card and contents (the EE Navigator);
  • a series shelf (Economics in 60 Seconds).

• One topic, six perspectives. Price elasticity of demand as understand, visualise, apply, calculate, interact and evaluate. These are keyboard tabs on a vertical rail, each stage independent.

• Inside the learning lab. A miniature of the platform's Predict a diagram lab, using its own questions and explanations for an indirect tax. There are five predictions with immediate feedback, then the model's own diagram.

• From concept to judgement (in the About chapter). The nine-step progression the resources are designed around, shown with a second real topic, a depreciation of the rupee.

• Interested in exploring more? An email prefilled with the resource the visitor chose.

Learning Studio (#/studio; the address is unchanged from the earlier name, Resource Studio):

• Items: nine pieces of original work, shown in part.
• Data model: each item has an exhibit form (visual, editorial, interactive or series), a disclosure level, a verification date and a public line saying what was checked.
• Workflow: in docs/studio/ADDING-RESOURCES.md.
• Visual system: tokens, typefaces, figures and exhibit forms are in docs/design-system.md.

Learning design (#/design) follows price elasticity of demand through the nine steps: concept, model, diagram, evidence, quantitative reasoning, stakeholder perspectives, analysis, evaluation and judgement.

Contact and collaboration (#/contact) offers six kinds of enquiry, each prefilling an email or WhatsApp message to the published address and number. Nothing is sent from the page.

The resource hub

Eleven categories, each a card with its own drawn mark, an audience label and a description. Above them sit an audience filter and a search that runs locally over the category names, their descriptions and their tags.

Every destination resolves through one configuration object near the top of the script:

const RESOURCE_LINKS = {
  master: "https://drive.google.com/drive/folders/189b-UnYEus-_8a300KLt6QiELwCN133W?usp=sharing",
  ppt: null, platforms: null, handbooks: null, questionBanks: null,
  assessments: null, revision: null, ia: null, ee: null,
  teacher: null, cases: null, tok: null
};
function resourceUrl(key){ return RESOURCE_LINKS[key] || RESOURCE_LINKS.master; }

To give a category its own Drive folder, replace its null with the folder URL. Nothing else changes. Until then the category falls back to the master library, so no card is ever a dead link. The self-test asserts the master URL exactly, checks that no second Drive URL has been introduced anywhere in the build, and checks that every link carries target="_blank" and rel="noopener noreferrer".

There is no Google Drive API, no OAuth, no credential, no API key, no backend and no database. The cards are links. Nothing is embedded, nothing is fetched, and the hub adds nothing to load time.

Resource links also appear where they are useful rather than only in the hub: a Go deeper card on the curriculum explorer, the concept pages and the syllabus tracker; a Classroom resources card on the teacher pages; and a Resources column in the footer.

TOK × Economics

A major section, and a careful one. Three registers are kept visibly apart on every page, each with its own badge and colour:

• Official IB TOK information: statements taken from IB sources, each carrying its source, the date it was verified and the scope of what that source settles.
• Arjun Agrawal TOK × Economics integration: this platform's framework. Not an IB requirement, and no part of DP Economics assessment asks for it.
• Teacher-created activities: classroom material written for this platform. Free to use and adapt. Not IB material.

The IA and examination lenses carry a fourth label, Optional TOK-informed thinking, and the page says in as many words that the Economics IA is assessed against its own criteria and the papers against their own mark schemes, neither of which mentions TOK. The self-test asserts that no page claims the IB requires an Economics TOK activity.

The TOK section links to the official IB material for anything administrative rather than reproducing it. What it teaches is the habit of asking what a claim rests on, which is the same skill the highest evaluation band in Economics describes.

Economic data

Data explorer, country profiles, markets and prices, sources and method (More › Economic data, #/data).

Reading a chart. Above the explorer's chart, a strip names what the series measures, where (the economies shown), when, its unit and its source. Beneath it: why the indicator matters, the largest one-year change in the chosen period, what might explain a movement (offered as possibilities to test, not conclusions) and what to be careful about, including that lines moving together do not show that one caused the other.

What it holds

• Six indicators for 25 economies, the world and the four World Bank income groups: GDP, GDP per capita, consumer-price inflation, population, the Gini index and CO₂ per person.
• Brent crude oil, the US 10-year Treasury yield, US unemployment, and 23 exchange rates against the US dollar.

Where it comes from

• Each series comes from the institution that publishes it: the World Bank, the U.S. Energy Information Administration, the Federal Reserve (H.15, and via FRED), the U.S. Bureau of Labor Statistics, and CDIAC-FF.
• They are copied from the open-data packages at github.com/datasets, retrieved on 1 October 2026, each with its licence.

How it is handled

• Nothing is estimated or filled in: a missing year stays missing.
• A computed or converted series says so:
  • GDP per capita is World Bank GDP ÷ World Bank population;
  • CO₂ is converted from tonnes of carbon.
• Two errors in the packages' own metadata were caught by checking values against known figures, and are recorded on Sources and method:
  • the inflation file is labelled as a price index;
  • the exchange-rate README describes four currencies quoted the other way round.

The charts

• One axis, with up to five economies in a validated, colour-blind-safe palette. A colour stays with its economy.
• Direct labels, a hover and keyboard crosshair, a text alternative and a table of the numbers.
• Beside each chart are the facts computed from the values shown, and questions an economist would ask that link to the lessons. They are questions, not explanations: the platform does not interpret a movement for you.

Loading

• The 230 KB file (assets/data/econ-data.js) loads only when a data page opens, and is then kept for offline use.
• Before it arrives, the page says what it holds. If it fails, the page says so and offers a retry, never a number.
• To refresh the data, run RETRIEVED=YYYY-MM-DD python3 tools/build/build_data.py --refresh.

Economists and ideas

Sixteen economists whose ideas the course still uses (More › Economists and ideas, #/ideas): Smith, Ricardo, Malthus, Marx, Marshall, Pigou, Keynes, Joan Robinson, Hayek, Friedman, Sen, Ostrom, Stiglitz, Banerjee, Duflo and Acemoglu.

Each profile covers:

• the problem they faced;
• three or four central ideas;
• the lasting contribution;
• what their models assume;
• influence;
• criticisms and limits;
• why it still matters;
• a common misreading;
• an inquiry question;
• key works and prizes, limited to those the writers could confirm;
• links into the lessons, glossary and diagram atlas, and to the thinkers they answered.

The editorial layer (this release) turns the profiles into intellectual history:

• An interactive timeline places each economist by school of thought, at the year of their major work, with the twelve economic events marked above. Hover, keyboard focus or a tap shows a preview (school, core idea, idea chain, major work) and draws who they built on and who they challenged. On a phone it becomes a dated list.
• Each profile opens with the central question, the core idea, the key contribution, the major work and an idea chain, labelled as interpretation. The written profile follows, regrouped: the problem, what they argued, who they built on and challenged, what came after, assumptions and critiques, modern relevance. Beside it: the course, related economic events, and connections for Cambridge and other A-level courses, university, and the Indian economy where there is a real one.
• How ideas connect is an influence map across ten traditions, with every relation also listed in text.
• Compare economists sets Keynes and Hayek, Friedman and Keynes, Smith and Marx, Pigou and Coase, and the capability approach against income measures side by side on seven dimensions, without declaring a winner.

The schools, chains, relations and comparisons were written for this release and reviewed for accuracy; the 22 review corrections are applied at build time from src/content/ideas/rt-v2.json.

Review. The profiles were reviewed adversarially before release. The review found no factual errors; it corrected one misattribution and one contested claim stated as fact, and added HL labels where the guide marks content HL only.

Economic events

Twelve episodes told as stories you can follow (More › Economic events, #/events): the Great Depression, the end of Bretton Woods, the 1970s stagflation, the oil shocks of 1973 and 1979, the Plaza Accord, India's 1991 crisis and reforms, the Asian financial crisis, the dot-com bubble, the global financial crisis, the euro crisis and COVID-19.

The event in seven stages. Each event opens with its story at a glance: before, trigger, transmission, crisis, policy response, consequences and long-term change, each drawn from the event's own text and each a link into its chapter. The chapter rail stays in view below the section tabs, marks the chapter being read and carries a reading-progress bar.

Scrolling. An event page scrolls from top to bottom by wheel, trackpad, touch, Page Down, Home and End, after a refresh, Back/Forward or a deep link. (Before this release the rail's observer called scrollIntoView on its sticky links, which pulled the window back up each time a chapter came into view; tests/events-scroll.mjs guards against its return.)

Each event has fourteen chapters: the world before; the trigger; a timeline you step through; the mechanism (opened only after the reader predicts it); the data; the policy response and who gained and lost; short- and long-run effects; how it travelled; competing readings; what the models explain and miss; what might have been done differently and a what-if; comparisons; connections (IB subtopics and key concepts, economics topics for Cambridge, university fields, themes for civil-services study, economists, causal pathways, glossary, Real World cases); practice, retrieval and reflection; and further reading.

Fact, interpretation, inference, controversy. Every statement about the world carries one of these four labels, so a well-documented fact never looks like one school's reading of it. Every number that does not come from the charted data is listed under How the numbers were checked, with its basis.

The data. Charts come from assets/data/history.js: US inflation from 1914, US growth from 1930, Shiller's monthly share prices, US and UK ten-year yields, Brent from 1987, US unemployment from 1941, exchange rates for nine currencies and a sixteen-economy panel of growth and inflation. Each series names its original source, the open-data package it came through, its licence and the retrieval date. Where a package was wrong it was corrected and the chart says so: US growth is computed from chained levels, because the package's growth column is shifted by a year. Series whose licence is unclear (Case-Shiller, VIX) were left out. Each chart marks the event on its time axis.

Compare and history. Compare events sets any two side by side (trigger, mechanism, response, short and long run, how it travelled) and asks what is similar, what is different and which mechanism dominates. Economics through time puts the twelve events and the sixteen economists on one axis.

Review. Four writers drafted the events from a shared brief, and a separate reviewer then checked dates, people, causal claims and every listed number. The review found no high-severity errors and made seven corrections (for example, the January 1991 emergency credit was the IMF's, not the World Bank's). The corrections are kept as patches in src/content/events/rt-events.json and applied at build time. tools/build/build_events.py validates every subtopic, diagram, glossary term, case, economist and chart series an event names.

Pathways and the Indian economy

Six ways in (More › Pathways, #/paths): IB DP Economics (the deepest, written to the guide), Cambridge International AS & A Level, school, university, civil services and the Indian economy, and Economics for everyone. Choosing a path changes what the home page suggests first and hides nothing.

What the pathways do not claim. The Cambridge and civil-services syllabi could not be checked against their official documents from this build, so those pathways carry no syllabus map, no topic numbers and no exam-format claims. Each says so on its card, and the events name economic topics rather than syllabus references for them. A self-test fails if that changes before the documents are checked.

The Indian economy hub (#/paths/the-indian-economy) gathers India's growth, inflation and inequality against the world, thirty-one dated policy cases in seven themes, the 1991 event and the economists who wrote about India. It states no current policy facts.

Why did this happen?

Eleven causal questions (Think › Why did this happen?): why inflation rose, unemployment rose, growth slowed, a currency depreciated, a central bank raised rates, a government spent more in a recession, a tax raised the price by less than the tax, output rose after a subsidy, a tariff helped producers but hurt consumers, supply-side reform takes years, and a market polluted too much.

How a question works

1. The student first predicts which way each variable moves.
2. They then step through the eleven stages: event, starting conditions, mechanism, who acts, the diagram (drawn from the atlas), what changes (with their prediction marked), short run, longer run, evidence (described, never quantified, linked to Real World cases), evaluation, and the other explanations, with how to tell them apart.

HL-only parts are labelled.

The full text of both features is in assets/data/ideas.js, loaded when a profile or pathway opens. Lists and search use a small resident index.

The course: thirty-one lessons

Every subtopic in the Economics guide (first assessment 2022) is a complete lesson at Course › Topics. The earlier topic dossier became the lesson, and its old address still works. Each lesson follows one loop: Learn → See → Try → Explain → Apply → Evaluate → Retrieve, then Connect.

• The big idea, why it matters and one real-world question.
• Before you start: the prerequisites, each a link to the lesson that teaches it.
• Objectives as things you can observe yourself doing: know, understand, apply, analyse and evaluate, plus calculate, draw, interpret and recommend where they apply. Where the guide sets a subtopic only at AO2, no AO3 practice question is attached to it.
• The core explanation in short blocks, with "because" chains, tables and a key-terms rail.
• The mechanism: cause → mechanism → immediate → secondary → long-run effect, and what it depends on.
• Model cards: purpose, assumptions, mechanism, prediction, application, limitations, real world, common misconception, exam use and a "Go deeper" note. Each is drawn with its plate.
• Try it: misconception checks with feedback, the calculations and the diagrams to explore.
• Explain it to me (simple, exam, deep) and Can I explain it?: you write first, then see what a complete explanation mentions.
• Apply it: the Real World cases filed under the subtopic, an inquiry lab, and an IA angle that is learning support, not a model commentary.
• Evaluate it: theory, evidence and interpretation kept apart; "Watch out", drawn from the misconception database; the professor layer; and the Economist's Lens.
• Exam connection: question types, command terms, diagram and calculation requirements, common errors, what strong answers do, and original practice questions. No claim is made about how often anything is examined.
• Retrieve it, then the seven Can I actually do this? ticks. The platform shows only the ticks that apply: no diagram, no "I can draw it".
• Connect: linked topics, key concepts and where each creates a tension, TOK, and a next step. What should I study next? reads only the written prerequisite links and your own ticks.
• Teach it adds a teacher layer: objective, starter, activity, discussion, formative check, extension and homework.

Foundation, Core and Deeper choose how much of the lesson to show. They describe the content, not the student.

Around the lessons:

• Unit pages.
• Glossary 2.0. 250 terms, each with plain English, the definition, why it matters, an example, related terms, the confusion, its topic and IB status, and an optional deeper note. Filter by letter, unit, topic, depth and status, and compare any terms.
• Command Terms 2.0. The IB's definition, then a teacher's guidance, each labelled.
• The misconception database. 51 entries, each with a student view and a teacher view.
• The Economist's Lens. Seventeen questions for any economic issue.
• The concept network. One concept, many worlds.
• Economics Intelligence. Each of the six real-world issues as a path from concept to exam.
• Paper guidance. Papers 1–3, with exercises on invented data, and economic writing.
• Checklists. Including "Can I actually do this?".
• Academic integrity. Includes "Before I submit" checklists.
• Revision mode. 15, 30 or 60 minutes, a unit, or the whole course.
• Snapshot mindmaps. The course, each unit and each subtopic on one page, printable on A4 landscape with a last-night version.

The course layer is data, not page fragments: assets/data/course.js carries its own version, its sources and its counts. How it was written, reviewed and patched, the source hierarchy, and a coverage matrix that shows every gap are in docs/content-architecture.md.

Masterclass and the teacher topic pack

A masterclass runs one subtopic end to end in the time the student actually has. Ten, twenty, thirty or sixty minutes, and a shorter session drops whole steps rather than shortening all of them: the ten-minute run does four things properly. The sequence never offers a step the subtopic cannot support, so a subtopic with no named calculation does not get a calculation step with nothing in it.

The topic pack builder lets a teacher choose a subtopic and what to include, and assembles the pack from material the platform already holds: objectives, conceptual core, activation prompts, the mindmap as an outline, key terms, a diagram checklist, the calculation with its common error, a five-minute retrieval starter, a real-world case, a TOK question, a misconception, exam questions and an exit ticket. Three versions print: teacher, student and blank. Items a subtopic does not have are left out rather than filled with generated material, and every pack states which parts are from the Economics guide and which are teacher-created. A self-test asserts that no pack claims to be an IB examination paper.

Economics in 60 Seconds

Timeless ideas. Everyday impact.

Ten original visual essays by Arjun Agrawal, supplied for inclusion and built in as a first-class learning layer inside Learn rather than as another menu item.

Issue	Title	Syllabus	Status
001	The Market for Lemons	2.10	Core
002	Comparative Advantage	4.1	Core
003	The Prisoner's Dilemma	2.11, 2.8	Enrichment
004	The Tragedy of the Commons	2.8	Core
005	Opportunity Cost	1.1	Core
006	The Power to Set Your Price	2.5, 2.11	Core
007	The Bill Someone Else Pays	2.8	Core
008	The Margin	1.1, 2.11	Enrichment
009	Incentives	2.3, 2.7	Core
010	The Sunk Cost Fallacy	2.4	Enrichment

Every cover is the issue's own first page, rendered from the supplied file, so the branding is the author's rather than a stock thumbnail.

Nothing is downloaded to show what an issue contains. The viewer holds the cover, the ten-page outline, the one-line idea, why it matters for IB Economics, the accuracy note, the questions and the links out to the mindmap, the real-world case and the diagram. The ten-page file opens in a new tab only when it is asked for, and when the assets folder is not beside the page the viewer says so and everything else still works.

Every issue was read page by page. A carousel has to be short, and short writing rounds corners. The note beside each one says where the shorthand would cost a mark if it were written that way in an examination, and it is specific rather than general:

• 001 already treats a degree as partly a costly signal and partly human capital, which is the correct position. The note adds that asymmetric information is one source of market failure rather than the only one, and separates adverse selection from moral hazard.
• 002 keeps the opportunity-cost rule and gains the condition it cannot fit: the gains require terms of trade between the two opportunity costs, and the model assumes constant costs and no transport.
• 003's payoff matrix was checked and is internally consistent. Betrayal is the dominant strategy for each player and mutual betrayal is the Nash equilibrium, worse for both than mutual silence. The note labels it a simplified one-round model and names what changes the result.
• 004 gains the distinction between a common pool resource and an open access resource, and Ostrom's evidence that commons are sometimes governed successfully rather than always collapsing.
• 005's film example is exact only if study really is the best alternative available. The note says so, and warns against summing every alternative.
• 006 gains the difference between inelastic demand and market power, which the carousel's own closing page also makes.
• 007 gains the formal definition behind its cover line, and the diagram consequence: MSC above MPC, MSB above MPB, and the welfare loss triangle between the free-market and optimal quantities.
• 008's "only one number matters" is read as "the total is not the relevant number", and the note separates sunk costs from fixed costs.
• 009's claim that the price rather than the rain summoned the drivers is qualified: both changed, and the honest claim is that the higher expected return raised quantity supplied. The colonial cobra bounty is flagged as an anecdote with thin documentation rather than as evidence.
• 010 gains the condition that a cost must actually be unrecoverable to be sunk, and notes that information gained from a past experience is not itself a sunk cost.

Original retrieval questions. Two or three per issue, each with one defensible answer. Teacher-created practice, labelled as such. No carousel is described as an official or endorsed IB resource anywhere, and a self-test scans every viewer for that claim.

Issue 024, "The Market Nobody Controls", is not in this build. It was named in the brief, the file was not supplied, and nothing was invented in its place.

The About page

An intellectual biography rather than a CV, read in under a minute:

• The cover: the name, five positions and one statement, with the portrait in its own cell.
• Credentials as data: education, higher education, IB Diploma Programme, subjects, mentorship and resource design.
• The journey as a map, ending in this platform, each stage opening what it taught.
• The classroom in two photographs.
• Six principles.
• The platform story: classroom, idea, resource, interaction, platform.

The earlier long prose is kept, folded, under The longer version. Every fact comes from the supplied profile.

• The portrait is never overlapped. The profile opens with three cells that cannot touch:

  • the text;
  • the portrait, in its own frame;
  • an economics panel of its own.

  No figure is drawn behind the photograph at any width, and a test measures the rendered boxes to hold it there.

• No current-employer framing anywhere. Fergusson College, BLISS International School and Symbiosis International School are presented as professional experience, with no institution described as a current post, on the About page, in the footer, in the page title, in the accessibility text or in the printed documents. A self-test scans every route and the footer for that framing and fails the build if it finds any.

• The photographs keep their own ratios. The portrait is used at 3:4, the two contextual images at 3:2 and 4:3. Nothing is forced into a square, zoomed into the face or stretched. Each declares its own aspect ratio and object position, and clicking any of them opens a larger view with its caption, its full alternative text, a close button and Escape support.

• The prose beside each image is about teaching, not about what the photograph contains: why a student is made to argue a side they did not choose, and why current affairs are where a model meets something that happened.

• Five positions: Economics educator, IB DP educator, teacher and mentor, academic resource creator, research-oriented educator.

• An interactive journey. Six stages, each opening to show what was taught, what it taught him, and what it changed about the way the platform works.

• UPSC CSE mentorship, stated as a fact, with no claim about ranks, selections or results.

• No superlative. No "best", no "number one", no "top IB educator", and a self-test asserts that none appears.

Your economic thinking

The homepage panel that reports learning signals was rebuilt rather than patched.

• No arbitrary score. A profile with no evidence behind it shows Not enough evidence yet, not a number. A dimension reports a figure only once three or more pieces of the student's own work sit behind it.
• Not an IB metric. Nothing on the panel is called an IB score, a predicted grade, official proficiency or an examiner score. It says what it is: learning signals from activity on this platform. The words "mastery index" no longer appear anywhere in the build.
• Plain language. "Decayed for time since retrieval" is gone. The panel says "Updated from your recent practice."
• A responsive grid, not a cramped radar. Twelve dimensions in five named groups, each with a short readable label, a one-line description and a progress bar. Verified at 360, 390, 430, 768, 1280 and 1440 pixels: no clipped label, no duplicate label, no cropped chart.
• It answers three questions and then offers one action: what is going well, what to work on next, what evidence supports that, and Practise this, which routes to the surface that trains the weakest dimension.

Coverage

The coverage dashboard is the instrument that makes a silent curriculum gap impossible. It lists every subtopic in the guide against every kind of material the platform can hold for it (mindmap, dossier, diagrams, calculations, questions, TOK, real world, curated video, teacher resource), and names the subtopics that have none of a given kind rather than leaving the reader to count. The table now counts real-world cases, curated videos and analysed question parts per subtopic as well, and prints the list of subtopics holding no case at all. The video coverage dashboard does the same for the video library, and reports nineteen of thirty-one subtopics rather than padding the library to look complete. It is organised by the IB syllabus, because the guide is the authoritative classification.

The mobile layer

A phone is not a narrow desktop, so the header is rebuilt rather than shrunk. Below 880px it becomes objects in a fixed order (logo, Resources, timer, profile, menu; the profile gives way below 430px), and the section navigation moves wholesale into a drawer grouped under Start, Explore, Practise, Progress, IA, Teacher and About. Every section in the build must appear in one of those groups; if one is ever added without being placed, it is appended to a More group rather than quietly becoming unreachable, and a self-test checks that.

Resource cards run three columns on desktop, two on tablet and one on phone. Wide tables and wide diagrams scroll inside their own containers rather than pushing the page sideways. Forms collapse to a single column on phones.

The release requirement is unexpected horizontal overflow = 0, verified across every view at 360, 375, 390, 430, 480, 600, 768, 900, 1024, 1280, 1440, 1600 and 1920 pixels: 1,937 view x width combinations in the release harness: together with checks for brand, navigation, timer, profile and menu collision, clipped header items, oversized tables, clipped diagrams and drawer overflow. Desktop layout is asserted separately so that nothing done for small screens can erode it.

Profiles and personalisation

On first launch the platform asks for a name, a level and (optionally) a target grade, exam session and usual study block. Several people can share one computer: each profile stores its own progress, notes, plan, mistakes, flashcards, drafts and mastery model, and profiles can be created, switched, backed up, restored and deleted from the profile drawer.

Personalisation sits beneath the platform identity. The workspace becomes "Aarav's Economics dashboard" inside Arjun Agrawal · IB DP Economics: the brand is not replaceable, and every printed or exported document carries the attribution.

Storage is versioned (AA_IB_ECONOMICS, schema v6) with migration logic, so future updates extend the structure rather than discarding existing progress.

Accuracy corrections in this release

Six corrections applied at source in this release, each checked against the Economics guide held in the project rather than against memory:

• Four macroeconomic objectives, plus government debt at HL, and external balance is not one of them. The guide's macroeconomic objectives in 3.3 are low unemployment, a low and stable rate of inflation, sustainable economic growth, an equitable distribution of income, and, at HL, a sustainable level of government (national) debt. External balance appears in the guide as a goal of monetary policy and of fiscal policy, in 3.5 and 3.6, and it is presented that way here.
• A sustainable level of government debt is filed under 3.3, not 3.6, and every node that names it now carries that code.
• Public goods: under-provision, not impossibility. The analytical claim is that the free-rider problem leads to under-provision relative to the social optimum. The absolute form, that private provision must be zero, is listed as a misconception to avoid. The merit good distinction is kept.
• Supply-side policy is not the only category that can raise output without raising the price level. It is described as the category that raises potential output, with its lags and its opportunity cost stated as the evaluation.
• The CPI is not "a basket nobody buys." The basket is built from what households actually spend on. The real limitations, substitution, quality change, new goods and differences between households, are given instead, and the basket builder makes the last one visible.
• Inflation, disinflation and deflation are kept apart. Disinflation is a fall in the rate of inflation with the price level still rising; deflation is a fall in the price level itself.

Alongside those, three corrections that current examiner guidance requires:

• The 800-word limit is not one of the three criterion F rubric requirements. It is a separate limit: moderators do not read beyond the 800th word. The IA command centre now marks the three rubric requirements as such and states plainly that the word limit is not one of them.
• Supplementary articles and statistics are not appropriate for a commentary. The commentary is assessed on the single submitted article, and the article checker now says so.
• IA diagrams must be adapted to the article and explained in the commentary. The diagram check asks what part of the article the diagram explains, which values and labels are adapted to it, and where that explanation appears, because a generic pasted plate cannot reach the top level of criterion A.

Four further corrections were found by the release audit rather than by the brief:

• A defect in the question factory. Every quiz entry in the topic dossiers passed its skill tag as an object rather than a string, so the dossier rendered [object Object] where the skill should have been and the skill routing silently failed. The factory now accepts either shape and two assertions guard it.
• The eighteenth node type was declared but unused. "Real-world case" now appears in seven maps, each pointing at a case the platform already holds in verified form.
• Two stale counts in the interface. A mindmap panel said "of the twelve node types" when there are eighteen, and two panels said "across the ten maps" when there are forty-two. Both now count what is there rather than repeating a number.
• The independent-platform statement is now set out in full, in the footer of every page and on the About page, and a self-test asserts it is present and that no IB logo or crest is loaded anywhere.

Found by this release's audit

14. A question option was the literal word "undefined". A macroeconomics checkpoint question carried it as a fourth option, which reached the screen inside timed sessions. It now carries a real distractor, and an assertion refuses any option that is a placeholder word.
15. Ten custom properties were referenced by the stylesheet and never defined. Rules written in earlier releases used shorter names for colours and typefaces than this palette gives them, so those declarations were silently dropped and the styling they intended never appeared. The ten names are now defined once as aliases, which repairs every call site rather than rewriting the rules one by one.
16. The Real World tabs were routed by the shell's tab dispatcher, which replaces the section renderer outright, so several new tabs rendered the wrong page while still rendering something valid. The renderer is now wired after the dispatcher, and the self-test checks that each tab shows the surface its label names rather than merely showing something.

A typographic pass. Em-dashes and en-dashes have been removed from the platform's own prose, which now uses hyphens, commas, colons and parentheses, and the supplied real-world collection was swept the same way. Three em-dashes remain, and deliberately: they are inside verbatim quotations from the Economics guide, where the punctuation belongs to the source rather than to this platform. Mathematical minus signs are untouched.

Alongside those, best fit is explained properly: a checklist is a diagnostic aid for finding the right descriptor, never a mark calculator, because it is not necessary for every aspect of a descriptor to be met for that mark to be awarded.

Contact

Contact details appear on the About page, the Tutorials page and in the contact drawer, and nowhere else:

• LinkedIn: https://www.linkedin.com/in/arjun-agrawal-560405310/
• WhatsApp: +91 7092844945 · https://wa.me/917092844945
• Instagram: @imarjun101 · https://instagram.com/imarjun101
• Email: [email protected]

No other account is listed, and no other personal information is exposed anywhere in the application.

Privacy

Student progress stays on the device unless the student exports it.

No accounts, no sign-in, no analytics, no tracking pixels, no cookies set by this site, and no data transmitted anywhere. Export, import and reset are all in Settings → Take it with you.

This holds for everything added in this release. The video player is created only on a click and loaded from YouTube's privacy-enhanced domain; watching contacts YouTube, and nothing from this platform is sent with it. An assignment link carries the chosen items and nothing else, so a teacher can set work without a server and without anything about a student leaving their device. The service worker caches this platform's own files and leaves every other origin alone.

Run it locally

Download or clone the repository and open index.html in any modern browser. Double-clicking the file is enough: there is no build step and no server required.

git clone https://github.com/<your-username>/arjun-agrawal-ib-economics.git
cd arjun-agrawal-ib-economics
# then open index.html

index.html is fully self-contained: the logo, portrait, diagrams and all logic are embedded. You can email that one file to a student and it will work on its own.

Two things that single file cannot carry with it live beside it: the Exam DNA corpus (assets/data/exam-dna.js) and the course layer (assets/data/course.js: the lessons, glossary, misconception database and snapshot maps). Open the page on its own and the pages that need either file say so plainly; the earlier dictionary and command-term pages stand in for the new ones. Keep the folder together and both work, including offline once the platform has been installed.

For the installable version, the service worker needs a real origin rather than a file:// path, so serve the folder over HTTP:

python3 -m http.server 8000
# then open http://localhost:8000

Publish it on GitHub Pages

1. Create the repository

Sign in to GitHub → New repository → name it arjun-agrawal-ib-economics → set it to Public → Create repository. Do not add a README; this repository already has one.

2. Upload the files

On the empty repository page choose uploading an existing file, then drag in the contents of this folder: index.html, README.md, LICENSE, .gitignore, 404.html, favicon.svg and the assets folder. Commit directly to the main branch.

Keep the folders. Drag the assets folder itself, not the files inside it. If the files arrive flattened at the top level, the page still opens, but the ten carousel PDFs, the app icons, the offline cache and the external Exam DNA and real-world corpora all miss, because the page asks for them at assets/carousels/, assets/icons/ and assets/data/. The expected layout is:

index.html  404.html  favicon.svg  manifest.webmanifest  service-worker.js  robots.txt
assets/data/        exam-dna.js, real-world.js, course.js
assets/icons/       app icons and apple-touch-icon.png
assets/carousels/   economics-in-60-seconds-001.pdf … 010.pdf
assets/brand/       source logo, monogram and portrait files (the page embeds its own copies)

The .github folder is hidden on macOS. In Finder press Cmd + Shift + . to show hidden files before dragging, or skip it, see step 3b.

3a. Enable Pages using the workflow (recommended)

Repository → Settings → Pages → under Build and deployment, set Source to GitHub Actions, and place pages.yml at .github/workflows/pages.yml. In this repository the file sits at the root, where GitHub does not read it, so the live site is published by step 3b (GitHub's built-in pages build and deployment from main).

3b. Or enable Pages without the workflow

Repository → Settings → Pages → Source: Deploy from a branch → Branch: main, Folder: / (root) → Save. This works even if you did not upload the .github folder.

4. Visit the live site

After a minute or two it will be available at:

https://<your-username>.github.io/arjun-agrawal-ib-economics/

The Pages settings screen shows the exact URL once deployment finishes.

A note on the repository name. The site is published under a path, not at a domain root: https://<your-username>.github.io/<repository-name>/. Every path inside the platform is relative, so it works whatever you call the repository, and this was tested by serving the package from a subdirectory: the service worker registers with the subpath as its scope, the manifest and icons resolve, all ten carousel PDFs resolve, no request escapes the subpath, and the site still works offline from there. You can rename the repository without editing anything.

5. Update it later

Edit or replace index.html in the repository (Add file → Upload files, keeping the same filename) and commit. The site updates automatically within a minute.

Optional: a custom domain: Settings → Pages → Custom domain, then add the CNAME record your domain provider asks for.

Licence

Two licences apply, and the distinction matters:

• The code is MIT licensed. Reuse, adapt and learn from it freely.
• The educational content and brand are not. The Arjun Agrawal name, logo, photograph, written explanations, teaching frames, questions, exemplars and diagram designs remain the author's property and are not licensed for reuse or redistribution.

Teachers and students are welcome to use the site in lessons and to link to it. Republishing the content as your own, or repackaging it commercially, is not permitted. See LICENSE for the full terms, or write to [email protected].

Tutorials

One-to-one IB DP Economics sessions: full syllabus, revision, Papers 1, 2 and 3, internal assessment and extended essay. Guidance, coaching and feedback; not ghostwriting.

All sessions are one-to-one and online. The terms, exactly as supplied:

Pathway	Fee	Sessions	What it is
Regular sessions	US$20	1 hour each	Any topic of your choice
IA support	US$100	6 × 1 hour	Guidance in writing the three commentaries
Economics EE supervisor support	US$150	agreed at enquiry	Academic supervision and guidance, not ghostwriting
Marathon Revision	US$300	20 × 1 hour	The entire syllabus before External Assessment
Exam Practice	US$200	12 × 1 hour	Questions only: Papers 1, 2 and 3

The Tutorials page (#/tutorials) sets these out as a narrative: why students get in touch, how sessions are taught, what an hour looks like, the five pathways (what each is, who it is for, what you do, what to expect), which pathway fits, how to get started, and eleven questions. Every pathway has its own enquiry link. There is no booking system: enquiries go by email or WhatsApp and are answered personally. The page carries no testimonials, ratings, outcomes or guarantees, and a self-test fails if any price, count or length changes.

To enquire, email [email protected] with whether you are SL or HL, your school, what you would like help with, and your preferred days and times.

Credits and disclaimer

Built by Arjun Agrawal: IB DP Facilitator in Economics, Business Management and Global Politics, Economics educator, UPSC CSE mentor and resource creator. Academic background: St. Stephen's College, University of Delhi. Professional experience includes Fergusson College, Pune, BLISS International School, Pune and Symbiosis International School, Pune.

This is an independent educational platform created by Arjun Agrawal. It is not an official IB platform and is not endorsed by the International Baccalaureate Organization. It is not affiliated with or produced by the International Baccalaureate Organization. "International Baccalaureate", "IB" and "Diploma Programme" are registered trademarks of the International Baccalaureate Organization. No IB logo or crest appears anywhere in this platform.

Contact: [email protected]