With COVID-19 there’s just so much data that’s out there. I think Andrew Cuomo has done a really good job of visual storytelling with data by condensing it down into understandable parts.
Most data analysts do not do this. When you have a complex dataset, what are some of the tips and tactics that analysts can use to sort of take that and tell a story around it that people understand.
With a complex data set, you can answer a whole lot of questions, and you can do a lot of really fun stuff. I think we’re all pretty aware of all the powerful machine-learning algorithms out there, a lot of these very, very complex statistical analyses. I find, honestly, that if you can’t tell a story, even if it’s a complex story, if you can’t tell it in a very simple way then what good is it?
You might have all of this data, you might have these really cool hierarchical clustering algorithms or whatever you call them – logistic regression of whatever – but if you can’t get down to and answer that question that your customer or client has then what good is it?
I think that’s where thinking of this in terms of visually telling a story with data helps me, and I think it helps other people to think about breaking that down. The first thing you need to do as a practitioner when you start having a conversation about a really complex topic is to make sure you find a common basis of understanding with your audience, with our client.
If you don’t have that common basis, that foundation from the get-go, you’ve lost them. It doesn’t matter how great your R2 is or this, that, the other. It’s over their head and you’ve lost them, and what good is it then?
So I really try to stay true to what the original questions were, and really think about how someone is going to use the data. We all fall under these traps of going off on tangents or going down rabbit holes. For me, analysis is really iterative.
Coming back to the questions, answering a few more questions, coming up with a few more findings, and then going back and iterating. Almost like a palimpsest.