Monday 8 April 2013

Cambridge Town Bumps

I started rowing just over a year ago and last summer competed in the Cambridge town bumps for the first time. Bumps racing is a form of rowing race where crews chase each other in single file, attempting to catch or 'bump' each other. In the Cambridge version racing takes place over 4 consecutive days. If one crew bumps another crew, they both drop out of that days race and swap starting positions for the next day. Postitions are maintained from one year to the next with clubs and crews working their way up or down the river over time.

This form of racing has given its name to a type of chart called a 'bumps chart', which shows changes in rank (or boat position in our case) over time.  This sort of chart has been blogged about and discussed in various places and there are countless examples all over the interwebs.  The 'official' chart for the Cambridge Town Bumps (town clubs as opposed to colleges) is a static image that looks like it's been generated using Excel.  This is the mens chart from last year taken from the CRA website:



It does the job, but it would be nice if there was an interactive version that allowed you to highlight individual crews or clubs, or maybe see the overall places gained and lost during the week.  There is an award each year for the best overall club called the John Jenner Trophy, but the results are stored in a nasty native html table near the bottom of the results page, which is difficult to read.  With this in mind, I thought I'd create my own version using Tableau Public.

The Mens and Ladies tabs contain the bumps charts for the respective divisions, but also show a summary table that shows how each club and crew has done during the week.  Selecting a club in the table will highlight all the crews from that club in the bumps chart.  This also works the other way around.

The John Jenner Trophy tab shows an overview of each clubs crews and performance as well as a chart showing the final positions.

I've given each club a colour that is close to the actual crew colours, which I think works ok.  The aim is allow users to be able to pick out a club more easily than if every line was the same colour.  I've tried to avoid a 'rainbow' effect by muting the colours slightly.

Anyway, it's a start.  Next step is to try and incorporate a few years worth of data to see what patterns emerge.




No comments:

Post a Comment