Here’s the problem: you want to see trends in your data, but you have so many lines it looks like spaghetti. Sometimes you can solve this by emphasizing the series of primary interest and de-emphasizing the others. But what if the subject is more about overall trends, and which series climbed and which fell?
Here’s an example using data on baby names (girls) registered in Ontario from 1917 to 2010. The series shown are the all time top 20 names plus the 20 top names in 1917 and the 20 top names in 2012 (I did this rather than trying use all of the many, many names that have been used). You can see certain information clearly from the line graph, such as which names peaked dramatically and subsequently fell off cliffs.
You cannot, however, readily answer the questions: “Which names became more/less popular over time?” and “What were the most popular names in 1917 and in 2010?”. A dynamic heat map easily reveals this.
A “dynamic heat map” is “dynamic” because it incorporates time as a dimension, along with the discrete category variable and its measure. By sorting the data on the measure’s value (frequency of name registration in a given year, in this example) in the most recent year, we end up with a chart that handles a lot of data and still presents it clearly.
Heat maps tend to be thought of as something to resort to if the data precludes the use of “better” i.e. simpler graphic forms. While that logically describes this case, I think that this application of heat maps is, rather than a second best tool, really the best tool for the job.
Here’s a link to an interactive version on Tableau Public: