My Photo

Your email address:


Powered by FeedBlitz

April 2018

Sun Mon Tue Wed Thu Fri Sat
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30          
Blog powered by Typepad

Become a Fan

« G2 is 4 | Main | Black Unicorn Hunting »

February 04, 2016

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

Rikin Khanna

Hi Jeff

I am a student studying in SMU. I attended your talk last month and loved it. I have been following your blog after that.

But I somewhat disagree with this post. I was thinking that maybe the definition of discovery is too narrow here. We may not be discovering new things by big data but we are definitely discovering new ways and methods of finding those new things. In order to support my argument, I would actually use an example from your own work. You told us that using Geo Spatial Analytics you were able to forecast the location of the asteroids which those astronomers couldn't with their current powerful computers. So finding this new method is a discovery in itself I believe. What do you think about this perspective?

Jeff Jonas

Placing this InfoWorld counter-point piece here: http://www.infoworld.com/article/3040708/analytics/data-visualization-showing-isnt-always-telling.html

And my counter-counter-point: I think you have missed my point while at the same time making a case for my point. For example, when you point out that “The heat map on the bottom right illustrates that default is more likely to occur on mortgages with more than 10 years employment length.” This visualization helps tell that story aka storytelling. But the question is how did you find this? Did you have a ‘big data visualization’ covering the walls containing a near endless combination of possible X and Y axis features? And with each, different time scales, geographical regions, and other often relevant dimensions? Did you pick out the one on the lower left because over the thousands of representations this one was more significant than the heat map showing mortgage defaults for those which considered age range, career, and marriage stability? Was this more interesting than the graphs on employee turn-over trends quarter to quarter in contrast to stock price?

My point is really: big pictures of all the data are pretty, but not that useful for finding novelty. Once something interesting is found then a bite-sized consumable picture (whether sourced from big data or not) makes for a useful way to explore for more understanding and storytelling.

I’d enjoy debating you on this one day. Albeit once we cleared up what we each meant it is likely we would be on similar page.

The comments to this entry are closed.