Skip to content
clariant @ arch : ~/clariant/projects/house-price-prediction
cd ~/projects

~/projects/house-price-prediction on ⎇ main [✓]❯cat case-study.md

shipped

House Price Prediction — Jabodetabek

End-to-end ML pipeline predicting house prices in the Jabodetabek metro area. Random-Forest model deployed behind a Flask API; I owned the data preprocessing stage.

House price prediction app — input form and predicted output
role
Data Preprocessing Contributor
period
Feb – May 2025
status
shipped
Python Scikit-Learn Pandas Next.js Flask

Context

Class project, team of 3. Real-estate pricing in Jabodetabek is extremely heterogeneous — same area, same square footage, prices differ wildly. The goal was a data-driven price estimator that gives a fast, objective baseline against a noisy scraped dataset.

What I built

  • Owned the data preprocessing stage: missing-value handling, duplicate removal, outlier treatment.
  • Cleaned the raw dataset from 3,500+ rows down to 2,397 high-quality rows without losing signal.
  • Worked end-to-end with the team through model tuning, evaluation, and deployment into a Flask API behind a Next.js front-end.

Stack

Python · Scikit-Learn · Pandas · Next.js · Flask · Git

Outcome

Final Random Forest regressor: R² ≈ 0.85. Shipped as a full interactive web app — the cleaning pass was the difference between a useless model and a useful one.

~/projects/house-price-prediction on ⎇ main [✓]❯cat links.md

cd ~/projects