← Challenge Hub

AI & Machine Learningintermediate

Crop Yield Forecasting

Predict crop yields from farm-household survey data covering eleven African countries.

Learn First

1
Introduction & Ethics
Introduction

Understand the data sources, formats, and the ethics of working with African data.
- Where the data comes from
- What questions it can answer
- Responsible use
2
Data Cleaning Fundamentals
Data Cleaning

Handle missing values, outliers, and inconsistencies.
```
import pandas as pd
df = pd.read_csv("data.csv")
df = df.dropna()
```
3
Exploratory Data Analysis
Exploratory Data Analysis

Visualize patterns across regions and time.
- Distributions
- Correlations
- Trends by district
4
Modeling & Evaluation
Modeling

Build and evaluate a simple model, then write your recommendations.
- Train/test split
- Baseline model
- Metrics that matter

Dataset

Title: Agricultural Survey of African Farm Households
Country: Multi-country (11 countries)
Source: Kaggle · survey microdata
License: CC BY-SA 4.0
Ethics note: Survey microdata is anonymised but household-level — do not attempt re-identification, and aggregate results when presenting to protect respondents.

Open dataset →Data dictionary Dataset details

Instructions

Goal

Forecast seasonal crop yields from weather and farm-survey features.

The data

Use the linked Agricultural Survey of African Farm Households dataset.

Steps

Open the starter notebook in Colab.
Join household, climate, and yield fields; handle missing values.
Engineer features and train a baseline regression model.
Evaluate (MAE / RMSE) and interpret the drivers of yield.

Deliverables

Your Colab notebook link
Key findings (3–5 bullets)

How to submit

Sign in and paste your work links in Submit your work for mentor review.

Rubric

Data Cleaning20 pts
Exploratory Analysis20 pts
Visualizations20 pts
Model / Evaluation20 pts
Report / Recommendations20 pts
Total100 pts