Tasty Data Goodies

Anchor Steam Brewery Tour

Last week Swivel ventured to the Anchor Steam Brewery for a tour and a little tasting. It was a nice break to get out of the office for the day.  Even though we were away from our desks, we couldn't help but soak up all the data tidbits Anchor Steam had to offer.

We learned that Anchor Steam likes to keep it local.  They brew and bottle all their beer in one San Francisco location.  They only use SF Municipal tap water and they ferment their beer at 60 degrees Fahrenheit, which is also the average temperature of the city by the bay.

The brewery's signature beer, Anchor Steam, makes up 70 percent of their sales, while their other six brews make up the remaining 30 percent.

And, being the data nerds that we are, here is a graph of the alcohol content in each kind of Anchor Steam brew.  Old Foghorn packs 8 to 10 percent. Woohoo.

Swivel goes to Washington

Last week was a very exciting week for Swivel. On Wednesday I showed Swivel at a panel "The Story Behind the Numbers: New Tools for Presenting Data," part of a publishing conference in Washington DC organized by The World Bank. I'm grateful to Eric Swanson of The World Bank for giving me the opportunity to present there. But that was just the beginning. In the short two days in DC, I met with several teams at The World Bank (thanks again to Eric Swanson), the IMF, the Pan American Health Organization (PAHO), and a small non-profit KEIonline.org. I had statisticians, economists, and computer systems people in these meetings, and while many of them had questions and concerns about topics such as data accuracy and caveats of data interpretation, overall they were open to the idea of putting data out for public debate and making it useful in the way Swivel is trying to do. Well, maybe "open to the idea" is a slight understatement: I got a hearty round of applause from a room of 40 people or so at the World Bank at the end of my presentation. One of the teams affiliated with The World Bank, AiDA, is already uploading data into Swivel as an Official Source. There will be more to come. Meeting these groups, showing Swivel, and listening to their comments and questions gave me a huge boost of enthusiasm that I tried to covey to the team back home, and at the same time highlighted all the many things we need to improve and problems we need to solve at Swivel. So, back to making Swivel better. 

Data Chief Joins Swivel

The day we launched Swivel, things were really hectic: the servers were running hot, the databases were melting down and none of us had slept for more than a few hours in the past 3 days.

Amid the chaos, we read a heartfelt blog that really stood out.  The author was Sara Wood.  In her blog entry, Sara talked about her vision for freeing the world's data and making it more fun and valuable - her dream for data was just like ours.  Importantly, Sara also illustrated a solid grasp of the pitfalls and stumbling blocks of unlocking value from loads and loads of data.

Her grasp for data is no surprise given her background.  Sara has spent the better part of the last decade at institutions working with some of the world's most important data: Harvard School of Public Health and the United Nations' World Health Organization.  She understands data and she understands the Web.

Just 3 months since we saw Sara's blog post, she has moved across the United States from Boston to join us here at Swivel HQ in San Francisco.  We are happy to announce that Sara Wood is now Vice President and Chief Data Officer here at Swivel.  Feel free to call her Data Chief...or just Sara is cool too.

Sara will be a key part of helping us free the world's data so curious people can discover and share new insights.

Welcome, Sara!

Inside Swivel: Graphing

Several of you have asked us about the graphing and visualization package we use to generate graphs at Swivel. Here's a bit of history about how we chose our graphing package.

In our quest for the right graphing package to use, we:

• Started with Gruff, a Ruby graphing library. It is well-designed, creates good-looking graphs and is easy to use. We realized though that Gruff at the time wouldn't scale as broadly as we wanted.  Gruff is fantastic and would be our first choice for smaller projects.
  
• Prototyped calling R, the free statistical package, via the command line interface (it has no formal API). R is a great tool and a lot of fun to work with. However, its presentation layer wasn't customizable enough for our needs, and it was also fairly slow, at least for the mass audience site we were building.
• Tried GNUplot, a package a lot of scientists are using, and rejected it as well because of limited presentation options, although it is very fast.
• Finally, we fell in love with Ploticus. It's a C package with an API. Brian wrote a Ruby API wrapper for it, rploticus, and Tao came up with a cool way to generate Ploticus templates in Ruby. We've been happy with what Ploticus gives us, it's got tons of options and graph types and is a very rich graphing package, and it's fast. You can tell we exposed only a small fraction of the various visualization options available in Ploticus, and as time goes by we'll be making better use of it and exposing more ways to see data.

If you truly want to geek out on visualizations, you can get an explosion of different ways to visualize data at IBM's Many Eyes, a project by the Visual Communication Lab, which is part of the Collaborative  User Experience research group at IBM's Watson Research Center. It's like Tufte on steroids. Their visualizations remind us of another graphing package we investigated called prefuse.

If you want any help with Ploticus or want to talk about other graphing packages drop us an email, we love to work on this stuff.