Hubway Bike Balance presents different aspects of bike station popularity. It shows which stations service most trips and how balanced incoming and outgoing trips were over the last months. Another aspect I explored is the correlation between closeness to public transport (MBTA stations) and bike station popularity.

The map consists of datasets provided by Hubway: A height model which I generalized to a 10×10 meter raster and used as input for a hillshade. Together with the water layer, this hillshade makes for a nice background which I overlaid with population density areas derived from Census blocks and tabular data. MBTA stations and lines form another layer. Using MMQGIS hub distance tool, I calculated distances between Hubway and MBTA stations for the correlation plot.

The Hubway station symbols are scaled by the average number of trips (incoming/2 + outgoing/2 per day) so more popular stations appear bigger. The symbol fill color shows how balanced incoming and outgoing trips are at a station (incoming – outgoing per day). Dark stations are more balanced than brighter ones.

Instead of labeling the stations, I decided to label the neighborhoods. The labels show a nice new cartographic feature in the current QGIS dev version: Transparent label buffers.

The map was created purely in QGIS – without any modifications in Inkscape or Gimp. The correlation plot was prepared in Calc and finished in Inkscape. For the final assembly, I arranged everything in Gimp, added the texts and improved the legend by adding explanations for symbol size.

12 comments
  1. Hi,
    Very nice map, but may I advise for a very minor change to the legend layout:

    Bike balance and population use a vertical legend layout, while trips per day is horizontal. I feel it would be more pleasing to have all of them vertical with legend headers like:
    – bike balance \n [bikes per day]
    – avergage trips \n [per day]
    – population \n [per ha]

    Also, this being the US with their weirdo imperial units, you might consider using acres instead of hectares…

    One more thing. Have you tried toning the water areas down a little bit? The contrast from the blueish-gray to the full white seems a little hard.

    But it has to be said: Very nice map production using our beloved FOSSGIS! Way to go!

    • Thanks a lot for your feedback Crischan! You raised some good points for another iteration. I guess the contrast issue you mentioned depends a lot on the viewer’s screen. It looks harmless here but I’ll try to tone it down a little. Wonder if I can get myself to use the imperial system though ;)

  2. Well, the old color management problem one encounters when stepping from the engeneering world of GIS into the artist’s world of creating maps… Me being in the GIS/Developing world (Mapbender3) I probably take anything as a screen as long as it outputs RGB and offers syntax highlighting… :) To be short: Might as well be that my screen blows the colors.

  3. I am just getting started with GIS, my background is in Public Health and i do research. Your maps are impressive. Any place you suggest for interesting base maps? I just installed QGIS today. i have been making maps with MapWindow. Simple things but more and more I want to create beautiful maps. Any suggestions would be appreciated

    • Hi Phillip! Some base maps such as Stamen’s Toner and OCM Landscape (available through QGIS OpenLayers plugin) can be quite nice to work with. But most are just to busy to add any more information to them. It’s more useful to get some base data such as elevation, coast lines, land cover, … to create your own background. A good start would be the data collection at Natural Earth especially if you are mapping bigger areas. Enjoy mapping!

    • My first idea would be to try to add this tile source to the OpenLayers plugin. I’ve never done it myself but there are examples that added Stamen’s tiles to the plugin so I guess it should work here too. Sorry I couldn’t be of more help.

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.