Bump charts are popular for exploring the changes in rank of a value over a time, place, or any other dimension relevant to the analysis. The name originated from a boat race where each boat tries to bump their boat and move up in their respective standing. They are very useful to compactly highlight trends and are also known as junk charts. A crisscross in a bump chart indicates one entity has surpassed other in absolute terms even when comparison is based on relative ranks.
Easy steps to create a bump chart
In order to create a bump chart, we need a ranked data set. As an example, Times Higher Education World Reputation Rankings data for the top 10 universities by reputation was used. The Times Higher Education World Reputation Rankings employ the world’s largest invitation-only academic opinion survey to provide the definitive list of the top 100 most powerful global university brands. A list of top 10 universities for the year 2017 was selected and then earlier ranks for years 2016 through 2012 were added.
Data was updated to indicate relative rank instead of overall rank out of 100:
A line chart with X-axis as column names, Y-axis as multiple dimensions, and color by institute would create a bump chart as shown in the following figure:
TIBCO Spotfire easily lets you highlight a line in bump charts to easily see institute name and changes in rank effectively. With the help of marking and tooltip, highlighting makes it apparent that Stanford has zigzagged a lot in terms of rank, and Berkeley has shown a dip in 2014.
It is also possible to use a parallel coordinate plot as a bump chart, but in that case, a calculation is required to give more weight to the highest rank.
A calculated column abs ([actual rank]-lowest rank-1) was used to calculate modified rank and this new rank was used in parallel coordinate plot to display bump chart as shown in the figure.
You can clearly determine Harvard has maintained its reputation at number 1 with Yale and Princeton having never gotten “bumped”.