Aventine Partners

Impressive Interactive Data Visualizations from NY Times

Despite the troubles in the newspaper industry, some -- notably the New York Times and Wall Street Journal -- have been exceptionally innovative in the area of data visualization.  Both have really raised their game in their widespread presentation of data in both the print and online versions.

A couple of recent visualizations from the New York Times team caught our attention as great examples of using motion charts to display change over time and interactivity to encourage exploring data.

An analysis entitled Turning a Corner?, uses a sequence of motion charts to display business cycle data and describe patterns found in past cycles and the current downturn.  What is interesting about this analysis is how nicely the analysis is explained and the story is told using nine 'frames' to the story.  This combination of i) an appropriate application of motion charts and ii) good story-boarding communicated the complex data, reasoning, and conclusion across in an very efficient fashion.

The Next Google?!

"I'm telling you, this company is getting hot . . ."

So began a pair of insistent emails from a friend and colleague about a company in the area I needed to check out.  It turns out there's a Google-killer right down the road in Champaign, home to the University of Illinois and the start-up website, Wolfram|Alpha (www.wolframalpha.com).

I say start-up "website" because this venture is the offspring of the 20+-year-old company Wolfram Research, which publishes the well-respected analytics software package Mathematica.

Though I am not a Mathematica user (like it, but I'm addicted to free), Wolfram|Alpha seems to be a web-based portal for accessing the Mathematica analytic engine and applying it to a large library of data.  The interface is designed to take natural language queries and return results in a format that will span a presumed range of intentions and levels of sophistication of the user, as well as allowing for further drill-downs where appropriate.  Hard to describe, so check out the easy-to-use site to get the drift.

The mission of the site/business is "making the world's knowledge computable".   Bold mission, that, and appropriately so for a Google-killer (not that I've seen these guys calling themselves that).

This seems like a great concept, and it makes sense as a true leap ahead in what a "search engine" does.  It will be interesting to see how it develops with more data and more users pointing it in unexpected directions.  

Limits to Data

Via the Freakonomics Blog I ran upon the departure announcement from Doug Bowman,  the head of the Google visual design team.  What is one of the driving forces behind his leaving Google?  Too much reliance on data to make decisions.

Here at Aventine we naturally highly value data.  But we also often talk about the limits of analytics as a science, which is to say experience and even instinct bring much to the process.  Typically experience makes the difference not so much between right and wrong or more or less accurate, but between more and less useful.

However, the Bowman post prompted some new questions, though he doesn't frame them this way much less answer them.  Is the problem with too much data and analysis:

1. It doesn't always lead to the "right" answer?
2. It is dehumanizing?
3. It is a nuisance?

My instinct (if that's allowed) is that it's a bit of all three, but especially one and two.  There is much that can't be measured and reduced to formulae, and part of our dignity as human beings lies in the fact that we do not mechanistically respond to events or stimuli like inert matter or mere animials. 

This is very relevant in the realm of predicting behavior, as we often do at Aventine, but even more so when dealing aesthtic questions and some the cognitive processes that good information visualization strives to enable.  In a funny way the power of our models keeps us employed, and the magnitude of the error terms keeps us human.

More on R and Alternative Statistical Analysis Packages

As a follow up to my prior post on the R statistical analysis tool, there have been a couple of interesting recent blog posts on the same topic, notably:

Comparison of data analysis packages: R, Mathlab, SciPy, Excel, SAS, SPSS, Stat, from from the AI and Social Science Blog, and

How Google and Facebook are Using R from the Data Evolution Blog.

Some quick takeaways:

• Google and Facebook are big users of R, which is the most popular analytics tool at Google.
• They use R for development, but not implementation of models, for which Python or C++ are the tools of choice. 
• The biggest debate seems to rage between users of R and SAS.  Many, many interesting takes on this question if the comments to the first post. 
• Ability to handle large data sets is one of the big differentiators and drivers of preferences.
• Some contend there is an age split, with younger users tending to prefer R and older users tending to prefer SAS.    

In any case, these make interesting reading for those facing these decisions.  Enjoy. 

Adults on Facebook

A recent study from the Pew Internet & American Life Project documented the rapidly increasing participation rates of older adults in social media such as Facebook and MySpace, as well as a range of other online activities.

Much of the information from the study is captured very nicely in the chart of online activity be age group:

(Pew Internet and American Life Project, Generations Online in 2009, Sydney Jones and Susannah Fox, January 28, 2009)

As a quick aside from the significance of the data presented here, this chart does a very nice job of communicating the degrees of penetration that various forms of online activity have made among the generations, from Gen Y to the G.I. Generation.

Bubble Charts, Good or Bad?

I like bubble charts.  It seems many others don't.  This is a quick review of bubble charts as a tool, with thoughts about why I like them, why others don't, and how best to use them.

At Aventine, our focus is marketing analytics, but we have a strong commitment to the related discipline of Information (or Data) Visualization, InfoVis for short.  By using good InfoVis techniques, we are able to create better graphical presentations of the analyses we do for clients.

I've watched with some interest recurring discussions in InfoVis circles about the use of bubble charts, including one recently on a site we like, Junk Charts, panning a graphic in the Financial Times using bubbles.  By and large, the InfoVis experts hate bubble charts, feeling people can't do a good job distinguishing differences of bubble size. 

The interesting, parallel development is that motion charts, which are just a type of animated bubble chart, are getting increasing attention thanks to Google and a remarkable presentation at a TED Conference in 2006.  These motion charts are getting traction despite the criticisms.