Museum of Almost Right: change an assumption, change the answer

Description and use

An interactive museum that shows how small assumption changes (sorting, data types, missing values, weighting) affect data analysis outcomes. The project is open source and publicly accessible.

Problem

Users need a way to explore how different assumptions affect data analysis outcomes without writing code or opening spreadsheets.

Outcome

Four synthetic exhibits with eight interpretations demonstrate that changing assumptions (e.g., numeric vs. string sorting, missing‑value handling) produces different, defensible results. The project is fully open source (MIT) and publicly accessible.

Workflow

• 1. Create Sanity documents for evidence, interpretations, and exhibits.
• 2. Build Astro frontend that queries Sanity and performs deterministic calculations in the browser.
• 3. Validate content with Studio schema and custom tests (25 tests, 25 passes).
• 4. Deploy to GitHub Pages and publish content via Studio.
• 5. Provide public URL for visitors to interact with exhibits and switch assumptions.

Evidence

The calculation keys select eight fixed functions. JSON is parsed as data, never executed.

Source

Both the main implementation context and a separate review context ran 25 tests with 25 passes.

Source