Nathaly Badra — Urban Data Analyst
nbadra.com · 2025–2026
This project analyzes the spatial distribution of conflict intensity and urban vulnerability across Israel, Lebanon, Gaza Strip, and the West Bank for the period 2021–2026, with a structural break at October 7, 2023.
The central contribution is the IVUB (Urban Conflict Vulnerability Index) — a composite index that adjusts conflict exposure by population, combining events, fatalities, war-period intensity, and war-period ratio at the regional level.
The project integrates four independent data sources, builds a reproducible pipeline in Google Colab and Python, and delivers interactive HTML maps, animated visualizations, and a bilingual dashboard.
Is there a spatial relationship between conflict intensity, shelter coverage deficit, and socioeconomic vulnerability in Israel and Lebanon — and how does the October 7, 2023 structural break reshape that distribution?
| Territory | Total Events (2021–2026) | Fatalities | % Post-Oct.7 |
|---|---|---|---|
| Israel | 15,484 | 1,993 | 72.5% |
| Lebanon | 27,747 | 5,505 | 76.6% |
| Gaza Strip | 34,486+ | — | 66.8% |
| West Bank | 41,336+ | — | 51.3% |
IVUB — Five most vulnerable regions:
| Rank | Region | Country | IVUB | Classification |
|---|---|---|---|---|
| 1 | Al Nabatieh | Lebanon | 1.000 | Muy alto |
| 2 | South | Lebanon | 0.660 | Alto |
| 3 | Baalbek-Hermel | Lebanon | 0.548 | Medio |
| 4 | HaDarom | Israel | 0.390 | Bajo |
| 5 | HaZafon + Golan | Israel | 0.318 | Bajo |
conflict_project/
├── README.md
├── conflict_urban_vulnerability_israel_lebanon_2021_2026.py
├── data_raw/
│ ├── Middle-East_aggregated_data_up_to-2026-03-07.csv
│ ├── worldpop_isr.tif
│ └── worldpop_lbn.tif
├── data_processed/
│ ├── acled_weekly_clean.csv
│ ├── acled_regions.csv
│ ├── worldpop_regions.csv
│ ├── ivub_regions_v2.csv
│ ├── ivub_districts_lebanon.csv
│ ├── ucdp_pts.csv
│ └── palestine_weekly_clean.csv
└── outputs/
├── acled_conflict_map_colab.html
├── acled_choropleth_v2.html
├── acled_ucdp_points_map.html
├── ivub_map_regions_v3.html
├── ivub_lebanon_districts.html
├── conflict_dashboard.html
├── conflict_timeseries_animated.html
├── conflict_timeseries_animated.gif
├── timeseries_four_territories.png
├── acled_timeseries.png
├── acled_regions_intensity.png
└── nota_metodologica_territorios_palestinos.txt
| Source | Dataset | Coverage | Access |
|---|---|---|---|
| ACLED | Middle East Aggregated Data | 2021-2026 · Israel, Lebanon, Palestine | Free registration |
| UCDP | GED v23.1 | 2021-2023 · Georeferenced events | Direct download |
| WorldPop | Population 2020 Constrained | Israel + Lebanon · 1km grid | Direct download |
| GADM | Administrative boundaries v4.1 | Israel + Lebanon (region + district level) | Direct download |
Territorial nomenclature note: The nomenclature used for "Gaza Strip" and "West Bank" follows the ACLED source convention and does not imply any political position by the author regarding the international status of any territory analyzed.
A composite index measuring territorial vulnerability to armed conflict, adjusted by population. Calculated at regional level for Israel and Lebanon.
Formula:
IVUB = 0.30 x (Events/100k)_norm
+ 0.30 x (Fatalities/100k)_norm
+ 0.25 x (War intensity/100k)_norm
+ 0.15 x (War ratio)_norm
Normalization: Min-Max per indicator with Winsorizing at p95 to control outliers.
Structural break: October 7, 2023. All indicators computed for two sub-periods: baseline (2021-Sep.2023) and war (Oct.2023-2026).
Methodological decisions:
- Golan Heights merged with HaZafon (Northern District) — both share the same ACLED centroid and are administered as a single unit in the source data
- Gaza Strip and West Bank excluded from IVUB due to: absence of comparable administrative polygons, unreliable population denominator post-displacement, and structural incomparability with Israeli/Lebanese governance contexts. Full rationale in
nota_metodologica_territorios_palestinos.txt
For Lebanon, the IVUB was extended to 26 districts using:
- Proportional disaggregation of ACLED regional data by WorldPop population weights
- UCDP GED v23.1 individual georeferenced events (spatial join to district polygons)
| File | Description |
|---|---|
acled_conflict_map_colab.html |
Heatmap + weekly conflict intensity by region |
acled_ucdp_points_map.html |
Individual events (UCDP) + ACLED intensity points |
acled_choropleth_v2.html |
Choropleth by country — events and fatalities layers |
ivub_map_regions_v3.html |
IVUB choropleth — Israel + Lebanon |
ivub_lebanon_districts.html |
IVUB choropleth — Lebanon at district level (26 units) |
| File | Description |
|---|---|
conflict_dashboard.html |
Bilingual (EN/ES) interactive dashboard — KPIs, timeseries, IVUB ranking, filters by country, event type and year |
| File | Description |
|---|---|
conflict_timeseries_animated.html |
Synchronized animated map + line chart with play/pause controls |
conflict_timeseries_animated.gif |
Exportable GIF for presentations |
| File | Description |
|---|---|
timeseries_four_territories.png |
Monthly events — Israel, Lebanon, Gaza Strip, West Bank |
acled_timeseries.png |
Weekly conflict intensity with structural break line |
acled_regions_intensity.png |
Events and fatalities by region |
Requirements: pandas · numpy · geopandas · folium · rasterio · matplotlib · pillow · requests
- Clone this repository
- Open
conflict_urban_vulnerability_israel_lebanon_2021_2026.pyin Google Colab - Mount Google Drive and set
BASEpath to yourconflict_project/folder - Execute Cell 0 — downloads and processes all data automatically on first run, reloads from Drive on subsequent runs
- Execute cells in order (0 through 36)
ACLED data requires manual upload — free registration at acleddata.com. All other sources download automatically.
Cell 0-4 → Setup, load, clean, regional aggregation
Cell 5-8 → GADM polygons + choropleth maps
Cell 9-11 → UCDP integration + individual events map
Cell 12-16 → WorldPop + IVUB construction + map
Cell 17-23 → Dashboard data + IVUB at district level (Lebanon)
Cell 24-27 → Dashboard + Drive upload + local download
Cell 28-33 → Timeseries animation + Palestinian Territories integration
Cell 34-36 → Regional charts + Folium map + dataset export
- ACLED data is aggregated at regional level (14 centroids for Israel + Lebanon). Individual event geolocation requires full API access
- IVUB is a relative index — values are meaningful for comparison between regions within this dataset, not as absolute vulnerability measures
- WorldPop 2020 population data may not reflect 2024-2026 demographic shifts driven by conflict-induced displacement, particularly in Lebanon's southern regions
- Gaza Strip and West Bank are included in conflict mapping and temporal analysis but excluded from the IVUB due to methodological constraints
- Golan Heights is merged with HaZafon (Northern District) in all IVUB calculations and maps
Badra, N. (2026). Conflict, Shelter & Urban Vulnerability:
Israel, Lebanon and Palestinian Territories (2021-2026).
Urban Data Analysis Project. nbadra.com
Data sources: ACLED, UCDP GED v23.1, WorldPop, GADM 4.1.
Code: MIT License · Data: per original source licenses · Analysis: CC BY 4.0 — Nathaly Badra