How to Document a Reproducible Football Chart

An exported football chart cannot explain how it was built. Without the question, source, cutoff, filters, population, metrics and units, another analyst can only reverse-engineer the image.

A reproducibility manifest preserves the minimum method needed to rebuild and challenge the result.

Document the Question

Record the exact analytical question and intended reader.

This prevents a chart created for exploration from being reused later as if it answered a recruitment or performance decision.

Record the Data Context

Include:

• source or product route;
• extraction or capture date;
• season or date range;
• competitions;
• last match or cutoff;
• known coverage limitations;
• provider-definition links where available.

Do not store credentials, cookies, personal session files or private identifiers in the manifest.

Define the Population

Record:

• position or role group;
• minimum minutes or matches;
• age range;
• included competitions;
• exclusions;
• comparison scope;
• population count when available.

The fair player comparison guide explains why each choice matters.

List Metrics and Units

For every field:

• current glossary label;
• definition link;
• total, per 90, percentage or contribution;
• transformation;
• direction;
• role in the chart.

For composite axes, include every component, formula and weight. See the composite-axis method.

Describe the Encoding

Record:

• chart type;
• X, Y and bubble-size fields;
• radar axis order;
• sort order;
• highlights;
• labels;
• thresholds or quadrant rules;
• aspect ratio;
• title and subtitle;
• style preset or relevant visual settings.

The manifest should distinguish quantitative encoding from editorial decoration.

Link Native and Annotated Files

If a chart is annotated externally, retain:

• native export;
• annotated derivative;
• annotation log;
• editor and date;
• crop or fading decisions;
• external calculations.

Use the annotation boundary.

Version the Configuration

Increment the manifest when:

• metric definitions change;
• filters or population change;
• a new period is used;
• the chart type or encoding changes;
• an error is corrected;
• the publication format changes interpretation.

Do not silently replace an old configuration if the published conclusion depended on it.

Minimal Manifest

Chart ID:
Question:
Owner and reviewer:
Created/updated:
Source and cutoff:
Period:
Population:
Eligibility:
Metrics and units:
Transformations:
Chart encoding:
Highlights:
Aspect ratio:
Native export:
Annotated derivative:
Limitations:
Changes:

Test Reproducibility

Ask another analyst to rebuild the chart using only the manifest.

Record:

• missing information;
• ambiguous definitions;
• unavailable settings;
• different results;
• corrections made.

A successful rebuild does not prove the interpretation, but it makes disagreement productive.

Connect the Workflow

The same record supports:

• a scouting shortlist;
• a report-ready chart;
• a weekly editorial series;
• future updates and corrections.

Rebuild This Chart From Its Visible Contract

The Jude Bellingham export already carries much of its own manifest: player, club, season, comparison role, same-league scope, per-90 mode, matches, minutes and last included game. The capture record adds the Playmaker metric preset and Broadcast classic style preset. Together, that visible and recorded contract lets a colleague challenge or rebuild the pizza without guessing what each slice means.

The separate configuration record adds what the image cannot: the exact route, player ID, expected preset, capture dimensions, output path, alt text and review status. If the production preset changes later, the old image remains interpretable and the difference can be documented instead of silently overwritten.

Try the handoff test. Give a colleague the chart and manifest but no verbal instructions. They should be able to reproduce the analytical state, identify any data refresh and explain why a value changed. If they can only recreate the colours, the documentation has preserved design rather than method. Reproducibility begins with the question and population; the PNG is merely one output of that contract.

Now imagine reopening the analysis six months later. The public preset may contain new metrics, a player may have changed club and the underlying dataset may have refreshed. A saved screenshot tells you what the old result looked like; the manifest tells you why. Preserve both. Rebuilding with today's defaults and calling it the same chart would erase the methodological change.

This is especially important when an annotated version circulates outside the team. Store the native export, derivative, alt text, caption and configuration under the same evidence ID. The identifier becomes a short path from a slide back to the analytical decision that produced it.

Document the Next Export

Create the chart in FBPlot and complete the manifest before writing the conclusion. Attach the record to the native export and keep both under the same chart ID. If another analyst cannot reproduce the population and encoding, the documentation is not finished.