Passion for Data Visualization

I discovered this on TED talks, and I completely agree that this is a true display of “passion for information visualization”.

The software used for the presentation is at http://www.gapminder.org and is also a google tool (http://tools.google.com/gapminder) It appears to be “bundled” with the global economic data, not sure if there is an open decoupled version that one can point to their data and play around. Although it seems flash-based and pulling data from static data sources (didn’t seem like rdbms, but I could be wrong) — but this would be a great way visualize OLAP data.

What’s interesting is the concept of a “play” button for the time dimension, which makes a great use of animation to see how different quantities (measures) change over time. It also manages the screen real-estate well to put different dimensions on X or Y-axis. But most of all, this truly exemplifies what data visualization is all about — it goes beyond the realm of charts and graphs that take a while to decipher, and rather tells a very clear, compelling, and visual story. Very impressive!

We’ve got charts and graphs to back us up, so f@#$ off!

Recently I had a potential partnership discussion to evaluate whether our predictive analytics technology could provide key insights from this potential partner’s (I’ll call them company XYZ) marketing database. Here’s how my conversation went with Mr. X, an exec at company XYZ, (which is a marketing technology company):

Me: “So, what are they key business pain points for your clients that we can analyze?”
Mr. X: “Well, you know, the usual stuff — marketing ROI, cut costs, increase sales, etc.”
Me: “ah.. yes, but can we delve a bit deeper? Where exactly clients’ marketing programs need help?”
Mr. X: “what do you mean?”
Me: “Well, are they more worried about increasing acquisition volume, or is it more about predicting high-LTV customers, or is it more about retention? I’m trying to get a sense of what is their #1 issue?”
Mr. X: “they don’t know.. it’s probably all of that stuff”
Me: “It’s important that we get a sense of priority, because otherwise we are talking about applying analytics without really knowing what we are trying to optimize”
Mr. X: “well, you are the analytics expert — you need to tell them what to analyze. They don’t think like you, worrying about success metrics, etc. They ask us to run marketing programs, and now we’d like to sell them some analytics. I can tell you what data we have on their marketing programs, now you tell me what kind of analytics you can provide me that I can sell.”
Me: “But we do need to understand their business objectives before determining what analytics is relevant enough so that they’ll pay for it”
Mr. X: “What I need from you is some screenshots — some charts and graphs that show what kind of analytics you can do — I’ll be more than happy to review that and tell you if we can work together”

Needless to say, I didn’t sense a true spirit of partnership here, but I did sense an attitude that I find more often than I’d like that analytics is all about producing charts and graphs that the user will somehow find useful.

Which, IMHO, is total BS!

But I can’t blame Mr. X too much because this is a pretty common perception of analytics in the marketplace. Recently I talked to a marketing exec who said –“everytime I meet with the analytics guys from our agency, they basically have this big ream of a powerpoint deck filled with one chart after another — and I don’t want to see all that stuff; all I want them to do is to tell me what relevant insight(s) did they find, and what course of marketing action would they recommend, and it’s like pulling teeth to get them to move beyond charts and graphs and talk about action.”

Look at the websites of any business intelligence software provider, or analytical software provider — and I will guarantee you that you will see a bunch of fancy charts and graphs, and dashboards with enough dials and speedometers to make you dizzy. Somewhere along the line, maybe we have forgotten that the purpose of analytics is to equip us with insights that enable better decision making.

So, first off, the type of analysis being done has to be aware of what type of decisions we are exactly expecting to improve; and second, the result of the analysis needs to be presented in a fashion that is “integrated” into the decision making. Maybe you list your recommended actions next to your charts and graphs, or maybe you somehow highlight the figures and trends that demand attention. My point — don’t leave it upto your user figure out the action based on the fancy charts and graphs, find out what decisions users are trying to make, and provide information that fills in that gap between analytics and actionable insight.