Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
Accepted Solutions
Well, I was GOING to explain that you shouldn't use a scatter chart for this. But since two other people already have, and since I've already built my example (see attached, though it isn't exactly what you're asking for), I'll play devil's advocate and explain why you SHOULD use a scatter chart. 🙂
Scatter charts CAN be good for noticing patterns that other types of graphs simply obscure, even when those patterns are not correlations between two variables.
Lets say we have a LARGE number of tasks, like 5000 of them. We want to see if there is a pattern to when they're due, like more tasks coming due towards the end of the month, for instance. We COULD create a scatter chart, one dot for each task, the due date along the bottom plus a random time to keep the dots from overlapping, and a meaningless random number along the Y axis to again prevent our dots from completely overlapping. Now we can clearly see our end of month pattern as rough stripes in our data. We'll probably see mostly white space on weekends. It might not be immediately clear what these patterns are, but we'll see that there are patterns.
In this case, it might be better to do it with a line chart, as detecting the pattern in the height of the line is as easy or easier. But with both of them sitting on top of each other, my eye seems to comprehend the scatter chart on a primal level faster. Line height has to be interpreted - OK, the higher the line, the larger the number of tasks due on that date, so peaks in the line correspond to clusters of tasks. Actually seeing those clusters is, for me at least, understood just that much faster.
OK, yes, I might not do it here. But I HAVE done it several live applications. In one case, we were interested in aged inventory per inventory location. The Y axis is the inventory location, given a numeric value, and randomly positioned within its "row". The X axis is the date and time that the item entered that inventory location. So in the image below, the further to the left the dot is, the longer it has been sitting in inventory. The orange dot is our target - we want to move material before it gets this old. There is a lot of obvious clustering in just this small sample of the data. And you can drag a box around a cluster of interest, and then see all of the details on other tabs, down to item IDs, hold codes, and that sort of thing. I could think of no better way to easily spot and then drill down to clusters of likely-related aged inventory.
