Logo Lanfrica
  • Home
  • Atlas
  • Insights
  • Docs
  • Sign in

© 2026 Lanfrica. All rights reserved. All copyrights of the resources shown on the Lanfrica website belong to the original copyright holders, unless explicitly stated otherwise.

BoldwinMax/Zambia-FRA-Maize-Price-Prediction

Domain:

agriculturesocioeconomic

Record type:

project
Creator:
Bol
Host:
Predicting Zambia's FRA maize floor price ahead of the annual announcement, using live World Bank & FRED data. Interactive Streamlit dashboard included. # 🌽 Zambia FRA Maize Floor Price - Predicting the 2026 Announcement Predicting the **Food Reserve Agency (FRA)** maize floor price (Kwacha per 50 kg bag) ahead of its annual announcement, using Zambian and global economic data from 2002–2025. **Headline result:** the model points to a 2026 floor price in the **mid-K350s** (indicative range ~K341–K371) - a modest, single-digit-percentage rise on the 2025 price of K340. > ⚠️ A small-sample, indicative model - see the caveats below. The FRA price is > ultimately a policy decision. ## 🔗 Live dashboard **➡️ Interactive dashboard** Move the sliders for the 2026 exchange rate, harvest, and inflation, and watch the predicted price update live. ## 📂 What's in this repo | File | What it is | |------|------------| | `Zambia-FRA-Maize-Price-Prediction.ipynb` | The full analysis — EDA, modelling, and verdict (renders directly on GitHub) | | `dashboard.py` | Interactive Streamlit dashboard (the "what-if" tool) | | `requirements.txt` | Python dependencies | ## 📊 Data sources | Feature | Source | Access | |---------|--------|--------| | Inflation (annual %) | World Bank API (`FP.CPI.TOTL.ZG`) | live | | Exchange rate (ZMW/US$) | World Bank API (`PA.NUS.FCRF`) | live | | Global maize price ($/t) | FRED API (`PMAIZMTUSDA`) | live | | FRA floor price | FRA announcements / academic records | embedded | | Maize harvest | ZamStats Crop Forecast Survey | embedded | | Fertiliser (DAP) | World Bank Pink Sheet | embedded | Live data is fetched at runtime, with embedded values as a fallback so the project always runs offline. All prices are in today's rebased kwacha (pre-2013 figures divided by 1,000). ## 🔬 Method A **Ridge regression** (regularised linear regression) on five features - previous year's price, exchange rate, inflation, maize harvest, and an election-year flag - evaluated with **leave-one-out cross-validation** (≈24 annual observations). The previous year's price and the exchange rate dominate; the FRA effectively …

Visit

github.com