KHobbs3/FEM-Niger-ETL

2025 Formative research ETL pipeline. # Niger FEM Survey — Analysis Pipeline This folder sits **outside** `niger_app/` and is the only place the raw PII dataset is touched. The Streamlit app reads only the pre-aggregated CSVs written to `niger_app/data/`. ``` / ├── pipeline/ ← this folder │ ├── config.py — paths + survey constants │ ├── utils.py — shared weighted aggregation helpers │ ├── agg_helpers.py — helpers for reading long-format CSVs in app pages │ ├── run_pipeline.py — master runner │ ├── etl_access.py — Health Access & Supply │ ├── etl_statements.py — Statement Agreement │ ├── etl_family_planning.py — Family Planning │ ├── etl_personality.py — Personality & Influencers │ └── etl_personas.py — Personas (k-modes clustering) | |__ validate_pipeline.py - Validates raw data and outputs │ └── niger_app/ ├── app.py ├── data/ ← pre-aggregated CSVs written here (safe to publish) │ ├── drivers_barriers.csv (already present) │ ├── radio_summary.csv (already present) │ ├── statement_labels.csv (already present) │ ├── access_stockouts.csv ← generated by pipeline │ ├── access_stockout_responses.csv ← generated by pipeline │ ├── access_travel.csv ← generated by pipeline │ ├── access_affordability.csv ← generated by pipeline │ ├── access_composite.csv ← generated by pipeline │ ├── statements_heatmap.csv ← generated by pipeline │ ├── fp_funnel.csv ← generated by pipeline │ ├── fp_timing.csv ← generated by pipeline │ ├── fp_methods.csv ← generated by pipeline │ ├── fp_reason_use.csv ← generated by pipeline │ ├── fp_intent.csv ← generated by pipeline │ ├── fp_nonuse_reasons.csv ← generated by pipeline │ ├── personality_life_goals.csv ← generated by pipeline │ ├── personality_ro …