🔫 DeepState Map Data

Every GeoJSON, in data folder, contains up-to-date Multipolygon, representing the russian-occupied territory of Ukraine.

Name format: deepstatemap_data_<update_date>.geojson

Frequency of updates: Daily, at 03:00 UTC.

To Do:
- Write a simple script using notebook code. ✅
- Set up Github Actions for daily updates. ✅
- Create a single compressed GeoJSON consolidating all data so far. ✅
- Write another script for Github Actions for daily updates of the above mentioned GeoJSON. ✅

📚 deepstate-map-data.geojson.gz -- Unified, Compressed Dataset

🟡 Important: All files within the data folder remain untouched, and continue to be daily updated.

The new compressed file contains all historical geomtries alongside their respective update dates, currently stored in the data folder.

Name format: deepstate-map-data.geojson.gz

Frequency of updates: Daily, at around 03:00 UTC.

Sample Data Structure:

id	date	geometry
0	2024-07-08	MULTIPOLYGON (((35.20146 45.52334, 35.31126 45...
1	2024-07-09	MULTIPOLYGON (((35.20146 45.52334, 35.31126 45...
...	...	...

* date represents date of update.

Accessing compressed data

If your application or tool does not support gzip-compressed GeoJSON files, below are several different ways I personally used to access/unzip the data.

Python (if compressed)

import geopandas as gpd
from io import StringIO
import gzip

with gzip.open("deepstate-map-data.geojson.gz", "rt", encoding="utf-8") as f:
    geojson_str = f.read()

gdf = gpd.read_file(StringIO(geojson_str))

print(gdf.head())

Python (if uncompressed)

import geopandas as gpd

file = "deepstate-map-data.geojson"
gdf = gpd.read_file(file)

print(gdf.head())

Linux Terminal (unzip)

gunzip -c deepstate-map-data.geojson.gz > deepstate-map-data.geojson

Windows 7-Zip archiver (unzip)

1. Right-click `deepstate-map-data.geojson.gz`
2. Select "Extract Here"
3. The file will decompress to `deepstate-map-data.geojson`

⚠️ Performance Warning

Since many features share locations and vary only slightly day to day, rendering all records at once (~400 as of 2025) using tools like geojson.io, can cause severe lag.

Try to filter date or use sample subsets before using less powerful rendering tools, if you experienced similar performance issues in the past.

Below: example of multiple layers stacking when loading full dataset without filters