EricNjiraini/paygo-cookstove-market-entry
Domaine:
socioeconomicenvironment and energyType de record:
project
Créateur:
Eri
Hôte:
PayGo clean cookstove market entry analysis — Kenya, Uganda, Tanzania, Ethiopia
# PayGo Clean Cookstove — Market Entry Analysis
**Kenya · Uganda · Tanzania · Ethiopia**
> A end-to-end portfolio project replicating the analytical workflow of a BI Manager at a PayGo energy company (e.g. M-KOPA, BURN, Sun King). Three interconnected modules — market sizing, credit risk scoring, and churn prediction — converging on a single market-entry recommendation.
---
## Project Narrative
You are the data analyst for a hypothetical PayGo cookstove company evaluating:
1. **Which country and sub-national zone** to enter first
2. **Who to extend credit to** at point of onboarding
3. **Who is likely to churn** and when — so field agents can intervene early
This mirrors real internal analytics at companies operating across East Africa's ~$1B PayGo clean energy market.
---
## Repository Structure
```
paygo-cookstove-market-entry/
├── notebooks/
│ ├── 01_data_collection.ipynb # API pulls, DHS download, data audit
│ ├── 02_country_comparison.ipynb # Module 1: macro market dimensions
│ ├── 03_market_entry_scorecard.ipynb # Module 1: country ranking + heat map
│ ├── 04_credit_risk_model.ipynb # Module 2: default probability model
│ ├── 05_churn_prediction.ipynb # Module 3: survival + 30-day classifier
│ └── 06_integrated_dashboard.ipynb # Ties all three modules together
├── data/
│ ├── raw/ # Downloaded source files (gitignored if large)
│ ├── processed/ # Cleaned, merged datasets
│ └── simulated/ # Synthetic customer ledger (see disclaimer)
├── outputs/
│ ├── figures/ # All charts and maps
│ ├── models/ # Serialised .pkl model files
│ └── tables/ # Scorecard CSVs, ranking outputs
├── docs/
│ ├── data_sources.md # Full provenance for every dataset
│ ├── methodology.md # Analytical decisions + assumptions
│ └── …