Nigerian House Rent Price Estimation System

A Streamlit app for predicting Nigerian house rent prices using machine learning. It includes interactive prediction, data exploration, location insights, and model performance dashboards.

Features

• Interactive price predictor based on property attributes
• Data analysis dashboard with histograms and comparisons
• Location insights (top/cheapest areas and stats)
• Model performance metrics and feature importance
• Multiple models: Random Forest, Gradient Boosting, Linear Regression

Quick Start

Prerequisites

• Python 3.9+ (3.11 recommended)
• pip

1. Create and activate a virtual environment (optional but recommended)

Windows (PowerShell):

python -m venv .venv
.\.venv\Scripts\Activate.ps1

macOS/Linux (bash):

python3 -m venv .venv
source .venv/bin/activate

2. Install dependencies

pip install -r requirements.txt

3. Ensure the dataset is present

• Place nigeria-rent.csv in the project root (same folder as app.py).
• Expected columns include: Price, Bedrooms, Bathrooms, Toilets, Location, Serviced, Newly Built, Furnished.

4. Run the app

streamlit run app.py

Then open the provided URL (usually http://localhost:8501).

Dataset

The app uses nigeria-rent.csv, which contains property listings with prices (NGN), attributes, and locations. The preprocessing pipeline:

• Cleans price strings and extracts numeric values
• Extracts numeric counts from text columns (bedrooms/bathrooms/toilets)
• Removes outliers and missing essential values
• Engineers features (total rooms, price per bedroom/room)
• Label‑encodes area names

Project Structure

app.py                 # Streamlit application UI
data_preprocessor.py   # Data cleaning and feature preparation
ml_models.py           # Model training, selection, and prediction logic
requirements.txt       # Python dependencies
nigeria-rent.csv       # Dataset (not tracked if large)
README.md              # Documentation

Technical Details

• Split: 80/20 train/test with train_test_split
• Models: RandomForestRegressor, GradientBoostingRegressor, LinearRegression
• Metrics: R^2, RMSE, MAE (reported per model)
• Feature importance: available for tree‑based models
• Caching: @st.cache_data speeds up reloads

The standalone training script in ml_models.py also supports optional hyperparameter tuning with GridSearchCV (cv=3).

Usage Guide

1. Price Predictor

• Set bedrooms, bathrooms, toilets, and amenities
• Select area; the app encodes it to match the model
• View the predicted rent and best model R^2

2. Data Analysis

• Price distribution histogram
• Average price by bedrooms/bathrooms
• Distribution of amenities (serviced/newly built/furnished)

3. Location Insights

• Top 10 most expensive and most affordable areas
• Area statistics table (average price and listing counts)

4. Model Performance

• Per‑model metrics (R^2, RMSE, MAE)
• Feature importance bar chart (if available)
• Best model highlight and brief recommendations

Train and Save a Model (optional)

Run the standalone trainer to save a model artifact:

python ml_models.py

This reads nigeria-rent.csv, trains models, may tune a tree‑based model, and saves rent_predictor_model.pkl with the estimator, scaler, and feature importance. The Streamlit app trains on the fly by default and does not require a pre‑saved model.

Requirements

streamlit==1.28.1
pandas==2.1.3
numpy==1.24.3
scikit-learn==1.3.2
plotly==5.17.0
seaborn==0.12.2
matplotlib==3.7.2

Troubleshooting

• Dataset not found: verify nigeria-rent.csv exists alongside app.py.
• Slow first run: models train on startup; subsequent runs are faster due to caching.
• Plotly axis updates: code uses update_xaxes/update_yaxes for compatibility.
• Info logs like "No runtime found, using MemoryCacheStorageManager" are harmless.

Contributing

1. Fork this repository
2. Create a feature branch
3. Make changes with tests/validation
4. Open a pull request

License

MIT License

Acknowledgments

• Dataset: Nigerian property rental data
• Libraries: Streamlit, scikit‑learn, Plotly, Pandas, NumPy

—

Built for the Nigerian real estate market.