SlugMeter — Newspaper Interview Talking Points

The Origin Story

Got a $48 parking ticket (P5 VC 22500A — parked within intersection). Deserved it.
Went to the AIMS (Automated Issuance Management System) portal to pay it.
The portal has zero authentication — you type in a citation number and get everything: location, time, violation, fine, vehicle make/model/color. No proof of ownership required.
Citation numbers are sequential: 26SC100001 = year 2026, device SC1, sequence 00001. Just increment the last digits to enumerate every ticket ever issued.
No rate limiting, no CAPTCHA, no bot detection. The system is wide open.
Wrote a scraper that evening, left it running overnight, had 200K+ citations by morning.

A real-time parking citation tracker for Santa Cruz with 218,000+ citations from 2019—2026.
Live map: new citations appear within minutes of being issued via WebSocket push.
Dashboard: filterable by device and time period, shows hourly/daily distributions, top violations, revenue.
Citation search: searchable, sortable archive of every citation with CSV bulk export.
Insights page: 6 analytical tabs — Revenue, Enforcement Patterns, Street Rankings, Equity Analysis, Anomaly Detection, Community contributions.
Risk Calculator: select any street and see a danger heatmap (day x hour), percentile rank vs all other streets.
Safe Parking Score: for any given day/hour, shows which locations are safest to park.

Data source: City of Santa Cruz public AIMS portal. Public record under the California Public Records Act (CPRA).
Citation number format: {2-digit year}SC{device number}{5-digit sequence}. Example: 26SC300042 = 42nd citation from device SC3 in 2026.
8 enforcement devices: SC1 through SC8. These are hardware units, not individual officers. SC8 alone accounts for ~25% of all tickets — likely a vehicle-mounted License Plate Recognition (LPR) system.
Scraping rate: ~8 seconds per citation (2 HTTP requests: search page + detail page). Slow, but steady and respectful of city infrastructure.
No PII stored: License plates and VINs are stripped before anything enters the database. Only public data: location, time, violation type, fine amount, vehicle make/model/color.
Geocoding: ~89% of citations are geocoded using Nominatim (free, open-source). The remaining ~11% reference parking lot space numbers (e.g., “LOT 3 SPACE #3042”) that can’t be placed on a map. They’re still in all aggregate stats, just not on the map.

| Layer | Tool |

|-------|------|

| Frontend | Next.js 15 (React) |

| Database | Supabase (PostgreSQL) |

| Real-time | WebSockets |

| Maps | MapLibre GL |

| Charts | Recharts |

| Geocoding | Nominatim (OpenStreetMap) |

| Hosting | Vercel |

| Styling | Tailwind CSS v4 |

SC8 is the busiest device (~25% of all tickets). Its volume and never-stopping pattern suggest it’s an LPR mounted on a vehicle, not a handheld unit.
Anomaly detection surfaces statistical outliers: days where a device wrote way more tickets than average (z-score > 2), “meter trap” locations that generate citations every single day, record days city-wide, and peak hours where enforcement concentrates.
Equity analysis: choropleth map overlaying citation density against census tract income data. Includes a scatter plot with Pearson correlation to show whether lower-income areas get disproportionately ticketed.
Community contribution: a Reddit user (camojorts) took the bulk CSV export, built a hex-binned density map using Folium with log10 scaling and per-weekday toggles. It’s embedded on the Insights page.

These are on the about page. Straightforward fixes:

Use non-sequential citation IDs — random or UUID-based IDs would make bulk enumeration infeasible.
Require proof of ownership — ask for the license plate or VIN alongside the citation number.
Implement rate limiting — even 10 lookups/minute per IP would make scraping orders of magnitude slower.
Add CAPTCHA or bot detection — prevents automated scripts from querying the system.
Separate public and private data — the API returns vehicle details in the same response as citation metadata. Split them.

None of these are exotic. Most modern citation portals already do several of them.