I have data like "5 people each completed 1 task in session 1, 4 people completed 2 tasks, 4 people completed 10 tasks, 1 person did 20." etc, and I want to visualize these. Histograms aren't really conveying the quantity of tasks completed, as the larger numbers (say 20 task) have a small number of people. What chart is the most effective way to visualize the three quantities (# people, # tasks, and # of people who completed # of tasks) in a way that is not visually confusing?
Histograms aren't really conveying the quantity of tasks completed...
One could argue the bars are conveying the data perfectly, appropriately showing very different values as very different sizes.
To the general issue of discerning values of different scales, some approaches:
- use a log scale with lines or points (but you have to special case zero values)
- use area, which is essentially a square root transformation in one dimension
- model the values and visualize the modeling parameters instead.
Whatever you choose for a visualization, it will be better at showing some things than others, so you often have to first decide which things are more important to see (e.g., absolute change, relative change, session vs. session, zero values, ...).