Pretty Data

Every Thursday afternoon I have a meeting with several data scientists and we discuss how things are going. I am going to end up missing the meeting today because I have another commitment. However it got me thinking about last week's meeting and the work I have done this week, as I still had to submit a progress report for everyone to review.

Last week we got to see a pretty graph in 3D space. The presenter rotated it around so you could tell it was really 3D and not just meant to look that way on the screen. Anyone that doesn't stare at data all day long for his or her job might have been impressed with the graphic. My first thought was, "So what? What does this graph tell me? All I can see are a bunch of dots that look like a vapor cloud."

Now I suppose I might be a little harsh, but seriously, if you had zoomed out, it would have looked like an oval dot on the screen. That got me thinking about what makes data useful. I may modify this in the future, but today I am thinking that useful data needs to tell a story and/or create an action. For instance, watching a graph of Facebook stock over the past few days indicates that the original investors got screwed, at least for now. This may cause others to see the stock as a good deal and buy some. Current investors may see this data and decide to unload it before it tanks any more (Facebook stock is actually up 2 points today . . . the last time I looked). When I look at a cloud of dots, I don't see a story, nor do I want to do anything. Well I may want to take a nap, but that doesn't count.

So the next time you see a bunch of data, ask yourself, "Is it useful? Does it tell a story? Does it make me want to do something? Or do I just want to take a nap?" If there is no story nor do you feel compelled to take action, then you can probably ignore it.