Windows on the Web: New Site for Crunching Numbers

By Bruce Kratofil
NABE Webmaster 

This is actually the second Windows on the Web column I turned in for this issue. The first column looked at the two new browser releases – Microsoft Internet Explorer 7 and Mozilla Firefox 2. It had some useful information, but it was boring. After all, how much more can be said about browsers? (I was even bored writing it.)

Then, about an hour after I shipped the article off to the editor, I happened to follow a link on the Freakonomics blog, and found something interesting to write about. It’s a new Web 2.0 site called Swivel.

Web 2.0

Before actually talking about Swivel, maybe we should look at that buzzword, Web 2.0, first.  While at the time we didn’t call it Web 1.0 (for the same reason that we didn’t call it World War I until World War II came along), the World Wide Web originally consisted of someone or something creating a site and filling it with their content. Then we, the users, came along and viewed that content.

In a Web 2.0 site only a platform or framework is provided. The users, who upload their creations to share, provide the content itself. Prime examples of Web 2.0 sites are YouTube, Flickr, Wikipedia, Blogger, Google Video and MySpace.

Swivel

Swivel wants to do for data what Wikipedia did for the encyclopedia. Their own description is “Swivel is a Web site for curious people to explore data.” Once you open up a free account at Swivel, you can upload data, create graphs from that data, or create graphs from other people’s data. This post at Freakonomics includes some comments by the creators. You also may want to try the tour they have at the site.

I wanted to try it out but didn’t really want to supply my own data. So I went over to FRED, the economic time series collection at the Federal Reserve Bank of St Louis. They were featuring the Canadian/US Dollar exchange rate, so I downloaded that. Reading a little bit of Swivel’s help files, I saw that they wanted spreadsheet data uploaded as CSV (comma separated value) files and only simple headings at the top of each data column. Uploading was done via a browser form within a webpage; once the data gets uploaded they ask you to describe it, cite the source, and tag it, which means coming up with a series of one-word descriptions, separated by spaces, for the data. I chose this list of tags: exchange rate dollar Canada US. You can see the results here.

Once you’ve created the graph, you can embed it in a blog or on another website. In fact, the graph below is actually coming from the Swivel site, while the rest of this page is coming to you from the NABE webserver.

What Is It Good For?

As professional business economists, most of us already have access to very sophisticated data sources and analytical tools. At this point, Swivel isn’t going to be a big addition to your toolset. On the other hand, it isn’t hard to think of lots of uses that an undergraduate student would have for this, or someone who, for some reason or other, can’t get the high-priced tools.

Where Swivel may really make a difference is in data sharing. While people aren’t going to be uploading their confidential company data (although there will be paid Swivel accounts that let you keep your data private), consider some of the uses in either academic or public policy discussions. Before, if someone published a controversial finding and another researcher wanted to examine the data looking for errors (or fudging), it wasn’t always easy to get a copy of the data set. Now, if someone wants to make a controversial claim, they can also say “Here’s my data – test it for yourself.” Arguments presented that way can carry more weight than those based on data that no one else can see.

There can also be problems with user-supplied data. Just like someone can write an erroneous Wikipedia article, someone will probably supply fake data. However, since you will be asked for your source, it is possible for others to verify. There probably will be some bad amateur econometrics being performed at Swivel, too. People will compare two nominal economic time series, find some high correlation, and immediately claim causality. That’s when it will be time for economists to use their comparative advantage and show how data analysis should really be done.

Swivel has only recently come out of closed beta-testing. The general public has only been able to use it about a week. As of Dec. 8, there were only about 1,100 data sets uploaded. However, if it exhibits anything like the growth of Wikipedia (which now has 1.5 million articles) it could turn into another Web 2.0 success.

About those Browsers

If you are still interested in browsers, you can see a longer review of Firefox 2 that I did here . If you want a comprehensive review of Microsoft Internet Explorer 7, a good one is here at the Windows Supersite.