Bike Sharing Demand: Two-Hourly and Daily Aggregations for Washington D.C. (2011-2012)

A two-year historical log from the Capital Bikeshare system in Washington D.C., USA, aggregated on two-hourly and daily bases. The dataset, created by Hadi Fanaee-T and Joao Gama, includes extracted weather and seasonal information to model rental demand. It was used in a tabular data benchmark for regression on categorical and numerical features.

Use Cases

• Forecasting bike rental demand based on temporal features like hour, day, and season.
• Analyzing the impact of weather conditions on urban mobility patterns.
• Modeling the correlation between holidays, working days, and transportation usage.
• Benchmarking regression algorithms on mixed categorical and numerical data.

Strengths

• Includes two full years of data (2011 and 2012) for seasonal analysis.
• Enriched with external data sources for weather, holidays, and seasonal settings.
• Explicitly records travel duration and position data, acting as a virtual sensor network for city mobility.

Limitations

• Row count and file size are unknown, limiting suitability assessment for large-scale projects.
• Column-level documentation beyond the listed attributes is absent; field semantics for unlisted columns must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

Capital Bikeshare system, Washington D.C., USA, with weather data from http://www.freemeteo.com.

Collection Method

Historical logs were aggregated on two-hourly and daily bases, then augmented with weather and seasonal information.

Time Range

2011 to 2012

Freshness

Data covers 2011-2012; last update date for the dataset file is unknown.

Geography

Washington D.C., USA