World GDP as a Moving Treemap

A silent, vertical data animation asks a counterfactual question: what if the size of each tile reflected its share of a changing whole? A fixed 100% canvas then morphs through 45 years while a large year counter, percentage labels, a progress line, and changing bottom commentary supply the story.

Why it works

• The first frame states the question and the visual rule together; the 1980 chart is already visible at 0:00.
• The tile area is the proof. Japan is 14.5% at 0:07.9, while China's area grows from 3.8% at 0:16.3 to 17.8% at 0:30.3.
• The fixed chart footprint, year counter, and 100% note let viewers compare changing shares without a new camera angle or explanation; the 2025 state holds from about 0:34.4 to the end.

Hook formula: Question + visual demonstration: “What if {{ENTITIES}} were sized by {{MEASURE}}?” Why it holds: The first image contains both a counterfactual and a concrete answer. A brief still period gives the viewer time to read the chart before the 45-year animation begins. The opening chart reports US 24.9% in 1980 while the main 1980 frame reports 26.0%; a new version should reconcile such discrepancies before rendering.

Retention map: re-hooks through the rising China tile and bottom-line change at roughly 0:19–0:27; payoff at about 0:34.4; no call to action; hard stop on the final state, with no apparent loop.

Mix: 100% full-screen motion graphic. There is one locked virtual camera and one chart scene.

Rhythm: No cuts. Annual values interpolate continuously across a fixed layout. The detector marks internal changes at 0:04.0, 0:05.2, 0:07.6, 0:10.0, 0:16.0, 0:22.2 and 0:30.0; these are useful cue points, not seven new shots. A 1980s sample appears at 0:07.9, 1990 at 0:10.3, 2000 at 0:16.3, 2015 at 0:26.7, and the 2025 hold at 0:34.4.

Rules this editor follows (inferred from frames):

• Keep the outer chart rectangle, top hierarchy, bottom commentary strip and small source footer in fixed positions.
• Animate the tile boundaries and labels together, so the values appear to drive the geometry.
• Update the year and progress marker as a redundant reading aid; change the bottom headline when the economic story changes.
• Reserve the last 7–8 seconds for the end state rather than racing through it.

Full shot list

• Look: Editorial data dashboard on near-black navy, approximately #07131D, with slightly lighter panels. Warm white type around #F3F5F6 and muted gray secondary type around #9AA9B3. Tile colors are persistent category cues: US deep blue ~#23669A, rest slate blue ~#294356, China crimson ~#AC3440, Japan plum ~#734479, Germany ochre ~#A17C2D, UK steel blue ~#455E88, France cyan blue ~#3483A1, India brown-orange ~#AD7134, South Korea teal ~#138C88. These are visual estimates from the JPEGs.
• Typography: Closest free match: Inter Bold/ExtraBold for title, year, labels and percentages; Inter Regular/Medium for subtitles and source text. Optional IBM Plex Mono Medium for the tiny edition tag. At 360×640, use about 22–26 px for the main title, 48–56 px for the year, 20–26 px for large tile percentages, 11–15 px for small tile labels, 13–16 px for the bottom headline, and 7–9 px for source/method lines. Fit labels to their tile; hide or abbreviate detail when a tile gets too small.
• Captions: None. The top title and changing bottom headline act as section cards. Use sentence case, white text, and selective yellow or mint emphasis rather than karaoke words.
• Graphics and overlays: A compact brand/series tag at top; horizontal progress ruler with small dots and yellow active segment; flags at tile tops; thin dark dividers; green-teal bordered Korea strip under the chart. The tile area fills a consistent rectangle of roughly 320×210 px within the 360×640 composition.
• B-roll: None.
• Transitions: One opening switch from question card to main chart at about 0:04; otherwise interpolate tile dimensions and labels without camera moves or hard cuts.
• Branding: Small series label and source footer; no end card.

Measured: No trustworthy speech. The apparent 80-wpm, nine-word line at 0:35.1 is low-confidence and conflicts with the uploader's explicit “No narration.” Pitch and pause measurements therefore do not describe a voice in this video.

Described: Silent-first, data-led explainer. The year, tile areas and headline provide the cadence.

Voice direction: Leave the production unvoiced. If {{PLATFORM}} requires narration, use an original calm, precise explainer archetype that reads only the framing question and one key comparison; keep the moving data and music readable without sound.

• Music: Original synthesized instrumental according to the post. The measured pulse is about 121 BPM (heuristic); it carries the entire 42 seconds under the graphic. Mood appears steady and restrained from the visual edit; exact instruments and melody are not established by the supplied sheets.
• Sound effects: No distinct effects are documented in the facts. Do not require whooshes for this format.
• Silence and space: No spoken pauses. Keep the ending music stable while the final chart is held.
• Mix targets: Reference is quiet at −22.1 LUFS integrated and −6.1 dBFS true peak. For a new export, retain clean headroom and check playback on a phone; master level may be adapted to its destination.

• Title: Counterfactual question plus explicit time span: “What if {{ENTITIES}} were sized by {{MEASURE}}? {{START_YEAR}}–{{END_YEAR}}.”
• Cover / thumbnail: The supplied cover shows a mid-series 1997 chart state, large year, white heading, multicolor treemap, and dark side extension. For a new version, choose a frame with a visually striking shift and make the measure legible at mobile size.
• Description, hashtags, pinned comment: The post explains the area rule, names the featured entities and residual, gives final values, cites the data series and license, and states interpolation and nominal-value caveats. No pinned comment is supplied. Use 2–3 relevant topic/format hashtags if the platform calls for them.
• Audience signals: 858 views and 5 likes at capture (5.8 likes per 1,000 views), on its upload day. No most-replayed map or comments were supplied; these figures do not establish a proven viral pattern.

There is zero spoken-word budget. Write only the text the viewer must read; keep the main title and method short enough to coexist with the chart.

[0:00–0:04 · QUESTION / STATIC START]
What if {{ENTITIES}} were sized by {{MEASURE}}?
{{START_YEAR}} · {{WHOLE}} = 100%
{{ONE_LINE_AREA_RULE}}

[0:04–0:10 · EARLY LEADERS]
{{EARLY_HEADLINE}}  (about 5–7 words)

[0:10–0:19 · TURN]
{{TURN_HEADLINE}}  (about 5–7 words)

[0:19–0:30 · LARGEST SHIFT]
{{RISER_HEADLINE}}  (about 5–7 words)

[0:30–0:34 · ARRIVAL]
{{END_YEAR}} · {{ARRIVAL_HEADLINE}}

[0:34–0:42 · PAYOFF HOLD]
{{FINAL_TAKEAWAY}}  (about 5–7 words)
{{SOURCE}} · {{MEASURE_DEFINITION}} · {{INTERPOLATION_NOTE}}

Script prompt (paste into any AI assistant with this blueprint):

For {{TOPIC}}, choose one measurable whole, 6–9 named parts, a residual, and annual values from {{DATA_SOURCE}} for {{START_YEAR}}–{{END_YEAR}}. Write a 42-second silent vertical-chart text plan using the six beats in section 10. Return the exact on-screen question, five short headlines, an area-rule line, labels, final takeaway, source note, and timing. Each headline is at most seven words. Make claims only where the data supports them. Identify the largest visual shift and schedule it for 0:19–0:30. Do not invent values or imitate the reference's wording.

Voice prompt: No voice. Communicate all necessary context with chart labels, a 100% rule, year, and short headline.

Visual prompts:

• Chart scene: Design a 9:16 dark-navy editorial infographic for {{TOPIC}}. Reserve top 18% for title/year/progress, middle 45% for a proportional 100% treemap, lower 16% for one changing takeaway, and bottom 8% for a small source/method note. Use consistent category colors, thin dividers, white labels, and compact icons only if licensed and meaningful. No people, footage, 3D camera, or photorealism.
• Cover: Use a legible chart frame at {{COVER_YEAR}} with a large year, measure name, a contrasting set of proportional tiles, and enough dark negative space for text.

Caption spec: No subtitles. Set Inter Bold title at about 24 px and changing takeaway at 14–16 px in a 360×640 design. Keep year around 52 px. Update text on chart beat boundaries, not every animated year.

Music prompt: Original/licensed minimal synthesized instrumental, steady ~120 BPM, restrained pulse, no vocals, no dramatic drop, clean ending after a 7–8 s final-state hold.

1. Build a checked table of {{YEAR}}, {{CATEGORY}}, {{VALUE}}, {{SHARE}}, and {{RESIDUAL}}. Compute every share against the same year's whole and enforce a 100% sum within rounding tolerance.
2. Choose a treemap layout that keeps categories consistently identifiable; define stable colors and label priority for narrow tiles.
3. Render a 360×640, 30-fps scene with a fixed dashboard, area-proportional tiles, year counter, progress ruler, and exact percentage labels.
4. Hold the opening year for 4 seconds. Animate annual data through about 0:34 using smooth interpolation; mark narrative headline changes near 0:10, 0:19 and 0:30.
5. Hold the final year until 0:42. Add source, methodology, and any estimation caveat in the footer and description.
6. Add an original/licensed instrumental bed, export H.264, inspect phone-size legibility and verify several annual frames numerically.

• The question and area rule are visible by 0:00; chart motion begins around 0:04.
• Every tile's rendered area matches its labeled share; named parts plus residual sum to 100% at every checked year.
• Hook and main chart show the same starting-year values; source vintage and interpolation are disclosed.
• One continuous graphic shot, no hard cuts, and a 7–8 s final hold.
• Small tiles remain identifiable on a 360 px wide phone display; labels do not overflow.
• No narration; music has a clean end, no clipping, and a loudness level suited to {{PLATFORM}}.
• Values and claims are checked against {{DATA_SOURCE}}, including any backcasts, revisions or estimates.

• The creator's signature (don't copy): The OneChart name/edition tag, exact palette, GDP country selection, flag arrangement, and original music.
• The transferable format: Counterfactual question, a fixed 100% proportional canvas, a time-lapse with a large year and progress ruler, data-led re-hook, and an unhurried final-state hold.
• Use your own dataset, text, identity and licensed sound.

_{Source: https://www.youtube.com/watch?v=Ct0woSYjoIg · measured with dissect 1.1.0 on 2026-10-07 · visual judgments marked as estimates or inferences}