🚲 London Bike Sharing Analysis
July 2025 · Python · Tableau · PostgreSQL · Data Visualisation
🔍 Project Overview
This project focused on understanding usage patterns and demand drivers of London bike-sharing data. Using a combination of Python and Tableau, I analysed user behaviour over time, weather conditions, and bike rental volumes to produce an interactive dashboard for deeper insights.
⚙️ Methodology
- Data Cleaning: Removed nulls and anomalies (e.g. unrealistic wind speeds or user counts).
- Feature Engineering: Extracted day-of-week, hour, and weather categories. Added rolling averages to smooth seasonal trends.
- Visualisation: Developed a Tableau dashboard with time series plots, weather vs usage heatmaps, and hourly demand bar charts.
- Statistical Modelling: Applied correlation and regression analysis to identify factors affecting rental volumes.
💡 Features
- Dashboard with drill-down by month, day, and hour
- Interactive filters for weather and holiday indicators
- Dynamic tooltips and hover explanations for accessibility
📉 Errors & Fixes
- Missing weather readings – interpolated using linear approximation.
- Misaligned timestamps between features – resolved by converting all time series into consistent hourly UTC format.
- Data export to Tableau not recognising types – fixed with enforced
float64 and datetime types in Python.
📈 Key Takeaways
- Weather conditions (especially temperature and humidity) strongly influence bike demand.
- Usage peaks around commuting hours and drops significantly on weekends and public holidays.
- Tableau is ideal for stakeholder-facing visualisation, while Python is excellent for preprocessing and modelling.
🔗 Repository & Dashboard
GitHub Repo: London Bike Sharing
Tableau Dashboard
← Back to Blog