Reconsider how pivot chart with multiple series works

[ElenaGorbatkova]

I think the current pivot chart uses an incorrect approach to visualising multiple data series.

1. We allow users to mix multiple different data types together and display them using a single legend:

2. We allow users to combine different types of data, for example, counts and sums, in the same chart.

3. When there is a large number of different data series, the dashboard will inevitably turn into visual clutter with broken layout.

4. Right now, displaying multiple series side by side creates the impression that we support drill-down (for example: selecting item kind, then viewing gender for each item, and then seeing the amount spent by each gender). In reality, this is not how it works, we are simply overlaying a second value of a completely different nature onto the same chart using irrelevant scale.

What’s needed:

• When a legend works with multiple parameters, it should visualise comparable data, for example, years or other time periods. This would allow users to compare year over year (or December to December), such as how many (Y) xyz responses a certain customer category (X) received over time. Currently, we allow visualising the number of respondents and the revenue amount on the same chart, which makes comparison unclear.

• If the legend works with multiple types of data, it should first visualise one type, and then allow switching to another type using an appropriate scale. In other words, the user could select as many series as they want, but we would provide a checkbox to switch between those series rather than displaying everything next to one another.