Occasionally, you will compare metrics on a time series and one value will be much larger than the other. For example, some artists receive many more plays on Last.fm than on YouTube. When this happens, the smallest metric can get squashed at the bottom of the graph:

The Normalise feature can give you a clearer view of this data. When you click the “Normalise” button, your y-axis scale changes to a percentage, instead of an actual metric value. The highest peak for your metric expands to 100%, and the lowest to 0%. This happens separately for each of the metrics on your graph, giving a clear outline of the different individual patterns of activity.

If you haven’t already, enable one or two individual metrics on your time series by clicking on the colour-key squares next to the artist name, and drag the date sliders to their maximum positions. Try to make it so that one metric is bigger than the other, like in the above screenshot.

Then, click the normalise button to get a better view of your metric patterns of activity:

You can normalise both daily and cumulative views. In the example below, there are several million more Last.fm plays than there are YouTube plays, and on the scale of the graph, the YouTube plays appear to have hardly increased at all:

However, by normalising the graph, you can see that in terms of percentage, YouTube plays were actually increasing faster than Last.fm plays at different periods throughout 2011:

Normalising a date selection

It is important to note that currently, each metric gets normalised across the entire time series – not for your selected dates. This means that if you normalise for a selection, sometimes your metric will not extend completely to the top and the bottom of the scale. Musicmetric will be including this feature at a later date.