Are your density maps communicating clearly?

Are your density maps communicating clearly?

Creating simple visualizations that communicate directly and clearly with users can be a tricky task. Recently, we took a bunch of public data from BIXI based in Montreal and created several analyses that will be added to this blog post as we publish them. BIXI provides open data regarding their bike rentals and stations across Montreal [1]

An important thought and debate surfaced during this development while trying to determine the size of density plots in the visualization; How do we determine whether out density maps are communicating the desired message?

Data analysis and visualization not only utilize scientific methods and principles, but artistic ones as well. This means that the intent of a message can be taken in many ways or few ways, depending on the artist's materialization of the message. So how can we make sure that our density map is communicating our desired message or a range of interpretations that fit within our desired message? Below we take three different size selections for our density analysis of BIXI stations added each year in Montreal and compare their pros, cons, and potential messaging.

Our initial desired message was relatively broad in order to give wiggle room for the analysis and final chosen messaging: Stations have mainly been added in "these" geographic areas.

Proposed Density Map #1 - Small Density Marks

Small Density Map


  • Allows users to better see number of stations actually added
  • More accurate "street-level" density map


  • Harder to make generalizations since individual points are more visible
  • Not too different from a standard dot map without a high density of dots

Some Messages Potentially Communicated

  • The impact of stations are limited to specific small radii as shown

Proposed Density Map #2 - Medium Density Marks

Medium Density Map


  • Gives generalized feel of where most stations have been added
  • More accurate "neighborhood-level" density map


  • Difficult to determine number of stations actually added

Some Messages Potentially Communicated

  • Several hotspots of stations have been added across Montreal
  • Specific neighborhoods have experienced significant station growth
  • BIXI has added many new stations outside of the high density east portion of Montreal to expand their reach

Proposed Density Map #3 - Large Density Marks

Large Density Map


  • General city view of where station growth has occurred
  • Communicates "city-level" hotspots of growing station availability


  • Not enough specificity to know where stations were added
  • Limited usefulness to know whether new stations are walkable distance from high traffic areas
  • Dilutes areas that had concentrated station growth (like Lachine in South Montreal)

Some Messages Potentially Communicated

  • Macro growth of station locations has occurred in the midwest and northeast portion of Montreal
  • Growth has been spread across nearly all parts of the city (which is false)

In the end we chose to communicate which neighborhoods have had the most stations added for each year, and therefore decided the second proposed density map. All three options could serve different messaging purposes and would have been useful depending on the audience we were trying to reach. We believed that by communicating a neighborhood level density map that we could keep the analysis useful for both the business and BIXI users. This is because the density area would be detailed enough to show walkable distances to the new stations but general enough to see trends of station growth in general city areas.

Below is a live side-by-side of all the different density maps scrolling by year. As the year changes, you can see how the density size has an impact on perception for years with a large number of new stations and low number of new stations. Meanwhile, we actually chose a density size between the small and medium settings proposed for the overall station density map. The reason for this is that we felt the medium sized density setting didn't provided an exaggerated view of bike stations. The setting between small and medium provided enough detail to know where stations actually were while not overstating their presence.

Our future analyses will include gallons/liters of gasoline saved, peak ride hours, most popular routes, where stations were added compared to route popularity, and more! Make sure to subscribe for notifications of when we release new blog posts.

[1] -


Leave a Comment

Your email address will not be published. Required fields are marked *

  • No products in the cart.