An interactive Machine Learning tool for predicting Berlin apartment rents (Warmmiete), achieving 84% R² accuracy.
As an architect based in Germany transitioning into data science, I developed this end-to-end Machine Learning project to decode the complex Berlin rental market.
Unlike standard tutorials, this project leverages architectural domain knowledge to perform deep feature engineering—quantifying the real value of "Altbau", specific "Kiez" (neighborhood) premiums, and amenities like fitted kitchens.
I developed an interactive GUI using ipywidgets. Users can input apartment specifications (Area, Year, Location, Amenities) and get a real-time rent estimation based on the AI model.
(Fig 1: The Interactive Rent Calculator)
Data analysis revealed the hidden value of architectural features:
- Location Premium: The model captures micro-location value. For example, Mitte is ~600€ more expensive than Marzahn for the same apartment.
- Altbau Effect: Old buildings (pre-1949) carry a significant "charm premium" (~230€), outweighing their potential energy inefficiency.
- Kitchen ROI: Installing a fitted kitchen increases the monthly rent valuation by ~117€ (net premium after adjusting for location), suggesting a payback period of ~34 months.
- Algorithm: Linear Regression (Optimized with One-Hot Encoding)
- R² Score: 84.06%
- MAE (Mean Absolute Error): ±220 €
Below is the comparison between Actual Rent vs. Predicted Rent. The alignment along the red dashed line indicates high prediction accuracy for typical apartments.
- Core: Python, Pandas, NumPy
- Visualization: Matplotlib, Seaborn
- Machine Learning: Scikit-Learn
- UI/Interaction: Ipywidgets
- Clone the repository.
- Install dependencies:
pip install -r requirements.txt
- Open Berlin_Rent_Project_publish.ipynb in Jupyter Notebook or PyCharm.
- Run all cells to train the model and launch the UI.
Data Source: Kaggle (Apartment rental offers in Germany)