iamTomson9/AI-Big-Data-Hackerthon-Botswana-Food-Inflation-prediction-Model-
Domaine:
socioeconomicType de record:
project
Créateur:
iam
Hôte:
# Forecasting Botswana Food-Price Inflation
Hackathon workspace for the IndabaX Botswana 2026 challenge, **Forecasting Botswana's Human Capital Under Global Economic Shocks**.
## Live Dashboard
**Judges can explore the hosted project dashboard here:**
### Open the Autobots Dev Botswana Food Inflation Dashboard
The dashboard provides a read-only view of the official 2024 best-model
forecast, chronological model-performance evidence, leaderboard feedback and
human-capital analysis. The official submission remains
`outputs/submission/best_model_predictions_2024.csv`.
The team must forecast Botswana's monthly food-price inflation (`FAO Item Code 23014`) for January-December 2024 using only information available through December 2023. The submission must compare a classical model with a deep-learning model, connect the forecast to at least two human-capital indicators, and translate the evidence into Botswana-specific policy advice.
## Start here
1. Read `docs/REPOSITORY_GUIDE.md` for a plain-language map of the repository.
2. Read `docs/codebase-and-development-handbook.docx` to understand every Python module and the full development workflow.
3. Read `Autobots Dev.md` and the files under `context/` in the listed order.
4. Assign team roles in `context/team-working-agreement.md`.
5. Obtain the organiser datasets and place them as described in `data/raw/README.md`.
6. Run the environment and raw-data audit described in `context/specs/01-project-setup-and-data-audit.md`.
7. Record every material choice in `docs/DECISION_LOG.md` while doing the work.
8. Do not train final models until data audit, cleaning, aggregation, and merge checks pass.
## Official Phase 1 deliverables
| Ref | Deliverable | Limit | Points |
| --- | --- | ---: | ---: |
| 1.1a | Best-model predictions CSV: `year_month,forecast`, exactly 12 rows | 5 MB | 20 |
| 1.1b | Feature Engineering Report | 4 pages | 12 |
| 1.1c | Model Comparison Report covering both models | 5 pages | 12 |
| 1.1d | Repr …