devscore 🌍
devscore is a Python package that computes a multi-dimensional Development Score using open geospatial, satellite, and survey-based indicators. It is inspired by cutting-edge development economics research from MIT, Oxford, and the World Bank.
It helps researchers, governments, and NGOs estimate local well-being and market access for any region in Africa.
Features
- Poverty Prediction: ML-based poverty estimation using satellite imagery, nightlights, and infrastructure data
- Market Access: Distance and travel time calculations to economic centers, markets, and key services
- Infrastructure Mapping: Density analysis of schools, hospitals, roads, and financial services from OSM
- Food Security Assessment: NDVI-based vegetation and agricultural productivity analysis
- Night-time Lights: Economic activity proxy using VIIRS satellite data
- Mobile Money Potential: Financial inclusion indicators
Installation
pip install devscore
Anaconda/Conda Installation (Recommended)
For better compatibility with geospatial dependencies, especially on Windows:
conda create -n devscore python=3.10 -y
conda activate devscore
conda install -c conda-forge geopandas rasterio osmnx scikit-learn xgboost -y
pip install h3 earthengine-api devscore
See CONDA_INSTALLATION.md for detailed conda setup instructions.
Install from Source
git clone https://github.com/idrissbado/devscore.git
cd devscore
pip install -e .
Dependencies
pip install numpy pandas geopandas rasterio earthengine-api osmnx scikit-learn xgboost requests h3 shapely
Quick Start
from devscore import compute_development_score
# Compute development score for a location
lat, lon = -1.2921, 36.8219 # Nairobi, Kenya
score = compute_development_score(lat, lon)
print(f"Development Score: {score['overall']:.3f}")
print(f" - Poverty Score: {score['poverty']:.3f}")
print(f" - Market Access: {score['market_access']:.3f}")
print(f" - Infrastructure: {score['infrastructure']:.3f}")
print(f" - Food Security: {score['food_security']:.3f}")
print(f" - Mobile Money: {score['mobile_money']:.3f}")
Using Dynamic Weights
Instead of fixed weights, use data-driven or expert-based methods:
from devscore.scoring.final_score import DevelopmentScoreCalculator
# Method 1: AHP (Analytical Hierarchy Process) - Expert-based
calculator = DevelopmentScoreCalculator(weight_method='ahp')
result = calculator.compute_development_score(lat, lon)
# Method 2: Entropy - Data-driven, based on variation
calculator = DevelopmentScoreCalculator(weight_method='entropy')
# Method 3: Auto - Robust average of multiple methods
calculator = DevelopmentScoreCalculator(weight_method='auto')
# Method 4: Custom weights
custom_weights = {
'poverty': 0.40,
'market_access': 0.25,
'infrastructure': 0.20,
'food_security': 0.10,
'mobile_money': 0.05
}
calculator = DevelopmentScoreCalculator(custom_weights=custom_weights)
See examples/dynamic_weights.py for comprehensive examples.
Data Sources
All data sources are open and freely accessible:
- Satellite Imagery: Sentinel-2 (AWS Open Data), Landsat 8/9
- Night-time Lights: NOAA VIIRS DNB
- Infrastructure: OpenStreetMap via OSMnx
- Population: WorldPop
- Wealth Data: DHS (Demographic and Health Surveys)
Methodology
The package implements a weighted aggregation of five key development indicators.
Weighting Methods:
The package supports multiple approaches to determine indicator weights:
- Fixed Weights (Legacy): Traditional research-based weights
- AHP (Analytical Hierarchy Process): Expert judgment-based (default)
- Entropy Method: Data-driven, based on information content
- PCA (Principal Component Analysis): Variance-based weighting
- CRITIC: Combines variability and correlation structure
- Auto: Robust average of multiple methods
Default AHP Formula:
Development Score = w₁ × Poverty + w₂ × Market Access +
w₃ × Infrastructure + w₄ × Food Security +
w₅ × Mobile Money
Where weights (w) are determined using AHP based on development economics research consensus. For fixed weights: w₁=0.35, w₂=0.20, w₃=0.20, w₄=0.15, w₅=0.10.
Each component is normalized to a 0-1 scale where 1 indicates highest development.
Components
-
Poverty Score: Machine learning model trained on DHS wealth index, using features like nightlights, NDVI, built-up area, population density, and road density
-
Market Access: Exponential decay function based on distance to markets, roads, hospitals, and economic centers
-
Infrastructure Index: Density of amenities (schools, clinics, markets, banks) per km²
-
Food Security: NDVI-based vegetation health indicator
-
Mobile Money: Financial inclusion proxy based on agent density and connectivity
Research Background
This package is inspired by academic research from:
- World Bank Development Economics (DEC)
- MIT Poverty Lab
- Oxford Centre for the Study of African Economies (CSAE)
- LSE International Growth Centre (IGC)
- Harvard Growth Lab
- Nature papers on satellite-based poverty prediction
📄 Read the full methodology paper: PAPER.md - A comprehensive 8000+ word article explaining the theoretical framework, algorithms, validation results, and practical applications.
📚 Additional Documentation:
Citation
If you use this package in your research, please cite:
@software{devscore2025,
title={DevScore: An Open-Source Framework for Multi-Dimensional Development Assessment Using Geospatial Data},
author={Bado, Idriss Olivier},
year={2025},
url={https://github.com/idrissbado/devscore}
}
License
MIT License
Contributing
Contributions are welcome! Please open an issue or submit a pull request.
Acknowledgments
Thanks to the open data community and organizations providing free access to geospatial data for development research.
Metadata
Release files for devscore 0.1.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| devscore-0.1.1.tar.gz | 42.8 kB | Without provenance | Details for devscore-0.1.1.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| devscore-0.1.1-py3-none-any.whl | Python 3 | none | any | Details for devscore-0.1.1-py3-none-any.whl |
Total release size: 93.0 kB
Release files / devscore-0.1.1.tar.gz
| Download URL | devscore-0.1.1.tar.gz |
|---|---|
| Size | 42.8 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
648c1aafc39425bd05952e266a3bb9901a8f70d8f13df6da12abbaa7686de7e0
|
|
BLAKE2b-256 checksum How to use checksums |
da5b3c76e92bd1b142670d33e6da01b5841ba15fe44cf964038d1a155a0641bd
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.0
|
Release files / devscore-0.1.1-py3-none-any.whl
| Download URL | devscore-0.1.1-py3-none-any.whl |
|---|---|
| Size | 50.2 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
65fc712b806c3491e91964855b27e845df1ebb2486f2685a5c9d47dea5e69a2e
|
|
BLAKE2b-256 checksum How to use checksums |
2a9dfc1113ea652fb4a53a72ed33497b817e24ffd7133cfbef899dc4539e5f2a
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.0
|
Security for devscore 0.1.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.