This analysis on bike sharing system is based on a dataset given in the data.csv file, which can be downloaded from https://drive.google.com/file/d/0B1rSzZNRiU8NYzlHSGNnaWh4Vjg/view?usp=sharing.

This file contains more than 1.8 million records of rides (each as a row). The columns are the unique bikeid, start and end station ids and start and stop time.

The most common purpose of analyzing bike sharing system is to learn the usage pattern at the bike stations and study their correlation with other features such as station location, weather, population, neighborhood,etc., as are provided in other online bikesharing competitions (e.g., www.bayareabikeshare.com/datachallenge). However, the current dataset does not provide any of these features. 

The focus of the current analysis is therefore, to find community groups among all bike stations and compare their daily/hourly usage so as to estimate the neighborhood and mainly usage of bikes. 

  The data is probably converted from New York Citi Bike data https://www.citibikenyc.com/system-data, the largest bike sharing system around the US, which has 332 stations as of May 2013.

The station names are hidden and locations are not given in the current csv file. 
  
  

By Yesu Feng
11/09/2014