About
Why SemtSkoru?
Plenty of open-data projects for Istanbul plot data on a map, but don't make it comparable. SemtSkoru distills air quality, green space access, traffic congestion, parking access, healthcare access, and public transit access into a single 0-100 score, so you can explore Istanbul's 39 districts with real data and compare two districts directly. It doesn't offer a preference-weighted "best district for you" recommendation — that's a deliberate scope decision; it shows each dimension on its own and leaves the interpretation to you.
How do we score?
Every dimension is a real measurement, mapped from raw data to 0-100 with a linear formula — never a guess or a subjective judgment:
- Air Quality
- The AQI (0-500, EPA standard) from İBB's live air-quality monitoring station is mapped to 0-100 by splitting the EPA's 6 categories (Good → Hazardous) into equal bands and inverting them. Lower AQI = higher score.
- Green Space
- A linearly decreasing score based on distance to the nearest park: 0 meters scores 100, 2,000 meters (about a 25-minute walk) or more scores 0.
- Transportation
- A linear score based on the district's average traffic speed: Istanbul's urban speed limit of 50 km/h or higher scores 100, a full stop (0 km/h) scores 0.
- Parking
- The average share of open capacity across the district's İBB İSPARK parking facilities (open capacity / total capacity, unweighted average per facility): completely empty (100% open) scores 100, completely full scores 0. This only covers facilities operated by İSPARK, is a small sample, and doesn't reflect on-street parking.
- Healthcare Access
- İBB publishes its "34-Minute Istanbul" Health Index at the neighborhood level; a district's average is the population-weighted average of its neighborhoods' health-access index values (neighborhoods with zero population are excluded from the average). İBB doesn't state that this index itself is on a 0-100 scale, so the MVP linearly scales the lowest and highest district averages actually observed across the real 39 districts to 0-100 (the lowest-scoring district gets 0, the highest gets 100).
- Public Transit Access
- İETT's citywide bus-stop dataset (15,486 stops) is matched to each district's boundary with a point-in-polygon test (since stops' district-code field doesn't offer a reliable name match). Rather than the raw stop count, we use density (stops per km²) — the stop count divided by the district's actual land area (in km², computed with a simple planar approximation scaled to the district's average latitude) — so a district doesn't get an unfair advantage just for being large. This density is then linearly scaled to 0-100 using the lowest and highest values actually observed across the real 39 districts (the lowest-scoring district gets 0, the highest gets 100) — not an official standard, but an honestly documented MVP assumption.
The overall score is a simple average of the dimensions that have data. If a district is missing data for a dimension (for example, districts without an air-quality monitoring station), that dimension is excluded from the average and the interface clearly shows "No data" — it is never filled in by guessing or interpolation.
Data sources
| Dimension | Source | Freshness | License |
|---|---|---|---|
| Air Quality | İBB Open Data Portal (live API) | Hourly (at the source) — synced daily | İBB Open Data License |
| Green Space | İBB Open Data Portal (GeoJSON) | Annual | İBB Open Data License |
| Transportation | İBB Open Data Portal (January 2025 CSV) | Historical — not live, clearly labeled in the UI | İBB Open Data License |
| Parking | İBB İSPARK Open Data Portal (live API) | Live | İBB Open Data License |
| Healthcare Access | İBB "34-Minute Istanbul" Health Index (GeoJSON) | Static — the source hasn't been updated since 2024, re-fetched weekly | İBB Open Data License |
| Public Transit Access | İETT Bus Stop Data (GeoJSON) | Static — infrastructure changes slowly, re-fetched weekly | İBB Open Data License |
| District boundaries | OpenStreetMap / Nominatim | Static (fetched once) | ODbL |
Open source
SemtSkoru is fully open source under the MIT license — all of the code, including the scoring formulas, is public on GitHub. If you find a bug or a data source changes, you're welcome to contribute.