When your car’s fuel drops below 10%, you need a quick way to find nearby gas stations. This project simulates a Houston-based car’s location, clusters gas stations using K-means, calculates distances with Haversine, and recommends the cheapest stations within 10 miles. Built with Jupyter and visualized in Tableau Public, it’s a practical solution inspired by my family’s automotive roots (40 years at Ford, Bosch project management).
- Data Fetching: Used Overpass QL to pull Houston gas station data in Jupyter.
- Clustering: Applied K-means (5 clusters) to group stations spatially using Python.
- Spatial Analysis: Calculated distances from the car to stations using Haversine (R=3959 miles).
- Recommendations: Filtered stations within 10 miles of the nearest cluster, sorted by price.
- Python: Pandas, Scikit-learn (K-means)
- Jupyter: For code development and data processing
- Tableau Public: My first exploration using Tableau.
- Tableau Hub and Spoke + Clusters: Haversine Display Gas Station Clusters
- Clone the repo:
git clone https://github.com/harriswe/FuelFinder-Houston - Open the notebook in Jupyter:
FuelFinder - Houston.ipynb(may require debug based on version). - Run all cells to generate the data and CSVs.
- View the Tableau above.
Inspired by my grandfather’s 40 years as a Ford testing engineer and my father’s work at Bosch, this project reflects my passion for automotive systems and data-driven solutions.
- Portfolio: Portfolio Website
- GitHub: GitHub Profile
- LinkedIn: LinkedIn Profile