FORRT Lighthouse

Interactive network explorer for replication and open science evidence.

Local setup with uv

# Install uv if you don't have it
curl -Lsf https://astral.sh/uv/install.sh | sh
# Clone / enter the project foldercd lighthouse
# Create venv and install dependencies
uv sync
# Run the development server
uv run flask --app app run --port 8080

Then open http://localhost:8080

macOS note: port 5000 is occupied by AirPlay Receiver — use 8080 or any other free port.

Project structure

lighthouse/
├── app.py # Flask server and API endpoints
├── import_xlsx.py # Script to convert Excel data → data/data.json
├── data/
│ └── data.json # Data read by the app at startup (generated, don't edit manually)
├── content/
│ └── about.md # Content of the About page (edit this to update it)
├── templates/
│ ├── index.html # Main app shell
│ └── about.html # About page wrapper (layout only)
├── static/
│ ├── css/
│ │ ├── style.css # All app styles
│ │ └── about.css # Styles specific to the About page
│ └── js/
│ └── app.js # All frontend logic (visualization, navigation, search)
└── FORRT_Lighthouse_Data.xlsx # Source data (edit this, then run import_xlsx.py)

How data flows

1. We are currently using a local copy of the data stored in: FORRT_Lighthouse_Data.xlsx, with three main sheets:
   • effects_review — one row per effect/phenomenon
   • papers_review — one row per paper linked to an effect
   • effects_wikipedia — Wikipedia entries linked to effects (entries marked as non-related in the validation column are excluded from the app)
2. Running import_xlsx.py converts the Excel into data/data.json
3. The Flask app loads data.json once at startup into memory
4. The frontend fetches data through the API as the user navigates

In the near future we will start reading directly from Google Sheets

Updating data

Edit FORRT_Lighthouse_Data.xlsx and run:

uv run python import_xlsx.py
# or with an explicit path:
uv run python import_xlsx.py path/to/FORRT_Lighthouse_Data.xlsx

This regenerates data/data.json. Restart the server to pick up the changes.

Key columns in effects_review

Column	Description
effect_name	Name of the effect
field	Broad academic field
subfield	Academic sub-field
cluster_a, cluster_b, cluster_c	Thematic groupings used for sub-navigation within a discipline
description	Plain-text description

Key columns in papers_review

Column	Description
effect_name	Links the paper to an effect
title	Paper title
doi	DOI
year	Publication year
apa_reference	APA citation
current_classification	Paper type (foundational, critique, meta_analysis, replication, reproduction)
summary	Plain-text summary

Key columns in effects_wikipedia

Column	Description
effect_name	Links the entry to an effect
wiki_title	Title of the Wikipedia article
validation	Relevance check — entries marked non-related are excluded from the app
wiki_url	URL of the Wikipedia article
year	Year the article was last reviewed

Editing the About page

The About page content lives in content/about.md — edit that file directly using Markdown. The app converts it to HTML automatically on each request; no restart needed.

Production

The app is configured for deployment on Render via render.yaml. Each push to main triggers an automatic redeploy.

To run in production manually:

uv run gunicorn app:app --bind 0.0.0.0:${PORT:-8080} --workers 4

Previous work

A previous iteration of this project is available on this repository.