Brian's Blog

16

In our last Conversation on Big Data we discussed the importance of integration, of designing data repositories so that facts are organized by a set of data elements that provide business context. Actually, good design is as important for a successful quantitative analytics program as it is to finding the perfect Italian suit, or pair of shoes, or furnishing your home. The IBM Center for The Business of Government and the Partnership for Public Service sponsor our Conversations on Big Data because we think there is value in communicating creative uses of quantitative analytics.

This blog takes that value proposition further – to show the value in creativity itself to quantities analytics. Whether you are building a sports car or an analytics program, good design should be your mantra. Some of you may be skeptical that there is a relationship between design as in a Ferrari Testarossa and design in data repositories and quantitative analytics. Yet there is a mingling of art and science in both. The polarization of “Art” and “Science” is a 20th century concept. Many of the great Renaissance masters were skilled at mathematics and physics as well as their art. Charles Darwin’s sketches are as artistic as a Picasso etching. So it should not be a surprise that good analytics require good design.

This is even more important for the output of quantitative analytics as for the data repositories that provide the input. The best way to sound a death knell for your analytics program is to present your findings in ways that fail to communicate effectively to your audience. Think back to the last time you were at a meeting where someone read out loud a set of slides with very little expression to their voice. How long was your interest held? What did you take away? Odds are your answers are “not long” and “not much.” Traditional reporting techniques are the equivalent of reading wordy slides in a monotone. A much more effective way to present your findings is to leverage data visualization.

Duke University defines data visualization as: “an umbrella term, usually covering both information and scientific visualization. This is a general way of talking about anything that converts data sources into a visual representation (like charts, graphs, maps, sometimes even just tables)”[i][1]

To use data visualization effectively, there are a few common guidelines[2]. The first one is know your audience. Are they familiar with your program and its goals? Do they understand how quantitative analytics works? What are they expecting and how do they assimilate information? Once you’ve defined whom you are presenting your findings to, the second guideline is create a framework. If your audience includes folks who have no knowledge of what your analytics program is about, or analytics at all, include some background. Even if your audience is educated about your program and analytics, you still must relate your visualization to the data in some way. Bubble charts, for example, can express data points according to many dimensions. The size of the bubble, its color, its position on the x and y axes can be leveraged to communicate, but your audience needs to know what those combinations mean. The third guideline is to tell a story with your data and your charts. Don’t approach your task as simply showing a group of findings backed up by numbers. Tell your audience the tale by starting at the beginning, proceeding through the middle and concluding with the punch line you want them to take away.

Once you have your story structured, you can think about what visualizations are most effective for each part. Many experts refer to the 3C’s of Visualization:

1.       Clarity: The ability to quickly understand what data the visual is displaying, and how it is displaying it.

2.       Connectivity: How well the visualization connects disparate data points.

3.       Concentration: How well the visualization brings certain (sets of) data points forward and focuses the viewer on them.[3]

Steve Beltz, the Assistant Director of the Recovery Operations Center, shared some examples of how data visualization is leveraged at the Recovery, Accountability and Transparency Board, which monitors how ARRA awards are distributed and spent. The Transparency is provided on the ARRA web site, where anyone can click on a map of the US and drill down into the awards granted to ensure that the money is being used as stated in the award.[4] Steve says that “the software … is actually representing the data in graphics … if you picture a landscape and you see these mountain ranges of different data points, the larger mound is something that gets talked about a lot … where these little teeny tiny mounds are only referenced once or twice … but those are actually what we are interested in, and we’ll zoom in on those and find out what’s being said in that area.”

A picture that lets someone instantly identify their particular area of interest, zoom in and expand or drill down into it, and gain knowledge about it is truly worth a thousand words – or slides!

Check out past Conversations on Big Data for more tips and insights on using big data to improve your organization’s mission effectiveness.



[1] http://guides.library.duke.edu/content.php?pid=355157&sid=2904817

[2] https://hbr.org/2013/04/the-three-elements-of-successf

[3] http://online-behavior.com/analytics/effective-data-visualization

[4] http://www.recovery.gov/arra/Transparency/RecoveryData/Pages/RecipientReportedDataMap.aspx?stateCode=FL&PROJSTATUS=NPC&AWARDTYPE=CGL



___

This article represents the views of the author only, and the information contained herein is of a general nature and is not intended to address the circumstances of any particular individual or entity. No one should act on such information without appropriate professional advice after a thorough examination of the particular situation.

 

Actions: E-mail | Permalink |