House Price Prediction — Aden City
A supervised machine learning regression project for estimating residential property prices in Aden using structured housing features and a reproducible scikit-learn pipeline.
0.984
R² Score
120
Properties
5
Features
01
The Problem
Property price estimation is influenced by multiple features including size, rooms, age, and location. This project explores a structured machine learning approach using locally relevant housing data from Aden.
02
The Solution
An end-to-end supervised regression workflow that loads and preprocesses housing data, encodes categorical information, trains a Linear Regression baseline, evaluates predictions, and visualizes actual versus predicted prices.
03
My Role
Sole developer. I developed the Python data-processing, regression, evaluation, prediction, and visualization workflow.
04
Architecture
Aden housing dataset
Data cleaning
Feature preparation
District encoding
80/20 train-test split
Linear Regression
Test-set predictions
MAE / MSE / RMSE / R² evaluation
Actual vs predicted visualization
05
Key Highlights
06
Results
R² ≈ 0.98395 on the project test split
MAE ≈ 3,836.66
RMSE ≈ 4,416.94
96 training records
24 test records
07
Challenges
Working with a relatively small dataset
Representing categorical district information
Avoiding overclaiming performance from a limited dataset
08
What I Learned
Strong test-set metrics do not automatically imply production-level real-estate valuation accuracy.
A baseline model provides a useful reference before comparing more complex algorithms.
Real-world deployment would require larger datasets, stronger categorical encoding, cross-validation, and external validation.
Technology