🚗💥 Days Since Ohio Crash

dayssinceohiocrash.com — Tracking every incident where a car has driven into a building in the Columbus, Ohio metro area.

Satirical. Sourced. Fully static. Costs nothing to run.

---

How It Works

1. A GitHub Actions cron job runs every morning at 8 AM ET
2. scripts/scraper.py searches Google News RSS, Reddit, and Columbus PD/Fire pages for matching incidents
3. New candidates are added to scripts/candidates.json and a pull request is opened for review
4. You review the PR, approve real incidents into incidents.json, delete false positives
5. Merging the PR triggers a Netlify deploy — the counter and feed update automatically

---

Setup

1. Deploy to Netlify

1. Go to app.netlify.com → Add new site → Import from GitHub
2. Select this repo (tonefox/days_since_ohio)
3. Build settings:
   • Build command: (leave blank — pure static)
   • Publish directory: .
4. Click Deploy

Netlify auto-deploys on every push to main.

2. Connect your domain

In Netlify → Site settings → Domain management:

• Add custom domain: dayssinceohiocrash.com
• Netlify will give you nameservers to set at name.com

3. Reddit API credentials (free)

1. Go to reddit.com/prefs/apps
2. Click Create another app
   • Type: script
   • Name: OhioCrashBot
   • Redirect URI: http://localhost
3. Copy the client ID (under the app name) and client secret

4. Add secrets to GitHub Actions

In your GitHub repo → Settings → Secrets and variables → Actions → New repository secret:

Secret name	Value
REDDIT_CLIENT_ID	your Reddit app client ID
REDDIT_CLIENT_SECRET	your Reddit app client secret

The scraper runs without Reddit credentials (falls back to Google News + PD/Fire only), but Reddit adds significant signal.

5. Generate icons (5 minutes)

1. Go to favicon.io/emoji-favicons
2. Search for "collision" or use 🚗
3. Download and place in the repo root:
   • icon-192.png
   • icon-512.png
   • favicon.ico

---

Daily Review Workflow

When the scraper opens a PR:

Option A — via GitHub (remote):

1. Open the PR → review scripts/candidates.json
2. Edit the file: fill in location, city, county, building_type, status, injuries, funny_note
3. Move approved objects into incidents.json (assign a sequential id)
4. Delete false positives from candidates.json
5. Merge the PR

Option B — locally (faster):

git pull
python scripts/approve.py    # interactive CLI walks you through each candidate
python scripts/generate_feed.py
git add incidents.json scripts/candidates.json feed.xml
git commit -m "approve: add incident #N"
git push

---

Adding an Incident Manually

Edit incidents.json directly:

{
  "id": 9,
  "date": "2026-04-11",
  "location": "Polaris Pkwy Starbucks",
  "city": "Columbus",
  "county": "Franklin",
  "building_type": "Restaurant",
  "headline": "Car drives into Polaris Starbucks; barista finishes order",
  "funny_note": "The car was not in the mobile order lane. The car is now in the lobby.",
  "source_url": "https://www.nbc4i.com/...",
  "source_name": "NBC4i Columbus",
  "injuries": "none",
  "status": "plywood_phase",
  "reddit_thread": null
}

status options: rebuilt | plywood_phase | condemned

Then run python scripts/generate_feed.py and push.

---

Project Structure

ohio-crash-watch/
├── index.html                  — main page (counter + feed)
├── about.html                  — satire disclaimer + tip submission
├── incidents.json              — approved incidents (source of truth)
├── feed.xml                    — RSS feed (auto-generated)
├── manifest.json               — PWA manifest
├── sw.js                       — service worker (offline support)
├── netlify.toml                — headers, cache rules, Netlify Forms
├── .github/
│   └── workflows/
│       └── daily-scrape.yml    — cron job + PR creation
scripts/
├── scraper.py                  — fetches candidates from 3 sources
├── approve.py                  — interactive CLI for reviewing candidates
├── generate_feed.py            — generates feed.xml from incidents.json
├── candidates.json             — pending review (never deployed)
└── requirements.txt

---

Tech Stack

Layer	Choice	Cost
Hosting	Netlify	Free
CI/CD	GitHub Actions	Free
Reddit API	PRAW + free OAuth app	Free
Google News RSS	feedparser	Free
Police/Fire scraping	requests + BeautifulSoup	Free
Domain	name.com	~$12/yr
Total		~$12/yr

---

Satire & Legal

This site is satirical. All incidents are real and sourced from public reporting. We do not publish personal information about drivers or anyone involved. See about.html for full methodology and inclusion criteria.