Analysing usage of HSL City bikes

Availability data from http://dev.hsl.fi/tmp/citybikes/

Points of interest:

• Visualizing availability rates
  • System wide - Peak hours / dates?
    • Effect of day of the week
    • Effect of the weather?
  • Per station - Best/worst time of the day to get a bike at your favourite stations?
  • Per area - No bikes around central railway station at 9am?
• Also reversely: analysing demand

Sample results

Uimastadion

Bike availability at "Uimastadion" bike station on 25.06.2017. Observations:

• People coming in for a morning swim around 8:00. And leaving around 10?
• More people coming between 12-16 for the afternoon sun?
• But also people leaving during 12-16 (notice the "zigzag")
• Everybody leaving at 16?
• Only two arrivals after 16! (Around 17:20 and 19:00)

Working / Residential

Dirrent "station profiles" are noticeable:

• Residential area
• Workplace area

Example

Two stations 500m apart from each other on 25.6.17:

• Meilahden sairaala
• Messeniuksen katu (in Taka-Töölö)

Notice the opposite availability / demand:

• Meilahti fills up in the morning and empties around 15
• Messeniuksenkatu empties in the morning and fills up again for the night

Legend

Line graphs uses aggregared data. Each datapoint is an aggregation for particular hour on that day.

• Max
• Mean
• Min

Instructions for running the r scripts

• Clone project
• Open and execute main.r
• Wait aroung 5-10 min for loading data from HSL
• Look at the plots created!

This function will draw a plot for each station every 2 seconds:

draw_plot_animate <- function(DF, stationIds) {
  for (i in stationIds) {
    draw_plot(DF, idFormat(i))
    Sys.sleep(2)
  }
}

Change the date on line 6 at main.r:

bikeDF <- getWebData("20170625")