Afghanistan Aid Targeting Replication Repo

This repo contains the replication code for the paper "Program targeting with machine learning and mobile phone data: Evidence from an anti-poverty intervention in Afghanistan" by Emily Aiken, Guadalupe Bedoya, Joshua Blumenstock, and Aidan Coville (2022). The repo is structured as follows:

Data

• data/survey.csv: Synthetic household survey data
• data/phone_features.csv: Synthetic featurized phone data, containing around a thousand features relating to mobile phone use
• data/interim_analysis_datasets: Most datasets saved during the analysis will be stored here

Scripts

• data/generate_synthetic_data.ipynb: Generates the synthetic survey and phone features datasets; adjust parameters for more or fewer observations, or to inject correlations between variables
• 0requirements.ipynb: Installs the required packages for the subsequent replication scripts 1-4
• 1survey.ipynb: Analysis of raw survey data and generation of additional survey-based outcomes (asset index, below poverty line)
• 2machinelearning.ipynb: Implementation and evaluation of machine learning models to predict survey outcomes from mobile phone features
• 3targeting.iynb: Targeting simulations to compare accuracy of targeting methods
• 4costs.ipynb: Calculation of targeting costs for PMT, CBT, and the phone-based approach

Results

• results/tables: All tables that are present in the paper are saved here in .csv format
• results/figures: All figures that are present in the paper are saved here in .png format
• results/simulations: Results from machine learning models (predictions, feature importances, the models themselves) are saved here

Dependencies

• python = 3.7.7
• numpy = 1.18.5
• pandas = 1.0.5
• matplotlib = 3.3.2
• seaborn = 0.10.1
• scikit-learn = 0.23.1
• scikit-misc = 0.1.3
• lightgbm = 2.3.0
• joblib = 0.15.1