A research tool for visually exploring transparency in foundations Statistics Donatarias consulta pública data. This is not a report, a paper, or a serious economic publication — it is a development tool for exploring foundations & Trust data visually.

Live demo: foundations_mx

This repo Mexico Foundations & Trusts review detailed data on donatios. We scraped an small portion of all of it and built an interactive treemap visualization where each rectangle's area is proportional to beneficiaries and color shows the exposure metric.

The repo includes scrapers, parsers, and a pipeline for writing custom LLM prompts to score and color foundations by any criteria. You write a prompt, the LLM scores to evaluate foundations activity, and the treemap colors accordingly. But you could write a different prompt for any question. See score.py for the prompt and scoring pipeline.

What "Foundations MX analisys" is NOT:

• It does not an auditor or an stadistical method specifically to analize money laundring or finantial risks on trusts.
• It does not account for network analisys u other method to identify donations flow.
• The scores are rough LLM estimates, not rigorous model analisys.

1. Fetch (fetch.py) not really need it if you download the foundation dataset.
2. Parse (parse_foundations.py, process.py) — Converts Excel into clean json (foundations.json) or Markdown files in markdown/.
3. Tabulate (make_csv.py) — Extracts structured fields into foundations.csv.
4. Score (score.py) — Sends each foundation's Markdown description to an LLM with a scoring rubric. Each foundation gets a score from 0-1 with a rationale. Results saved to scores.json. Fork this to write your own prompts.
5. Build site data (build_site_data.py) — Merges CSV stats and AI exposure scores into a compact site/data.json for the frontend.
6. Website (site/index.html) — Interactive treemap visualization with color layer for risk exposure.

uv sync

Requires an OpenRouter API key in .env:

OPENROUTER_API_KEY=your_key_here

Download the foundations dataset This dataset contains more than 10,000 files, so I recommend you to copy only a subset of these into a folder named "data/2024" in the same level as the folder "site". You can look in foundations.csv, the column "ref" has the file names used in this visualization.

Or in your agent code type "setup" to manage the previous steps.

# Generate CSV summary (foundations.csv) from files inside of data/2024
uv run python make_csv.py

# Generate Markdown from Excel (this will make a folder named markdown)
uv run python process.py

# Generate foundations.json file
uv run python parse_foundations.py

# Score AI exposure (uses OpenRouter API)
uv run python score.py

# Build website data
uv run python build_site_data.py

# Serve the site locally
cd site && python -m http.server 8000

# Test some foundations with a specific model
uv run python score.py --test [RFC_1,...] --model [MODEL_NAME]

# You can run a subset with:
uv run python score.py --start [INDEX] --end [INDEX]
or
uv run python score.py --add [RFC_1,...]

Inside foundations.json you can consult the RFC and foundations names.