Tasty Data Goodies

Data Visualization in the Public Sector: Open Call for Evidence

The organization, Oxford Consultants for Social Inclusion (OCSI), is currently reviewing best practices and methods of data visualization for public sector decision-makers and researchers.

The project aims to understand the potential for visualizations to analyze and communicate data. They are currently looking for good examples of visualization that have: made an impact and helped communicate findings to decision makers; and helped researchers understand the story behind the data.

The project and its resulting website will be an excellent opportunity to showcase your own examples of good visualization, and help contribute to standards of best practice within the research community. If you have examples, OCSI wants to hear from you!

We would love to hear your stories too. When have data visualizations helped you with research or decision-making, in your personal life or in business? What kinds of visualizations work, or don't work? Post your comments below.

(Via infosthetics).

Who's got your data?

These days, the average person has so much information online: banking and credit card accounts, business data, social security numbers, medical prescriptions, you name it. Since the trend of putting personal information online isn't slowing down anytime soon, should we be asking the question: who's got your data? And more importantly, is it safe?

Reports of data breaches in the US increased 47 percent in 2008 from the year before, exposing the personal records of at least 35.7 million Americans, according to the Identity Resource Theft Center. Most breaches were electronic (82.3%), while just a handful were on good old fashioned paper (17.7%). The graphs below show the number of breaches and the number of records breached by sector in 2008:

According to the study, 240 breaches were made in the business sector last year. As a company that offers businesses a place to store and explore their data privately online, we're very concerned with data security. You can rest assured, Swivel's taking every possible measure to keep your information secure.

Tracking Swivel Metrics (Part 2)

In the spirit of self-education, we're sharing the numbers we track at Swivel.  In this second post on the topic, let's take a look at users. The first post is available here.

To understand user behavior, the best practices include watching three key metrics:

• Recency - how long ago was the last time that the user interacted with your site?
• Frequency - how many times has the user come back to your site?
• Monetary value - how much money has the user spent on you? 

Recency is the most informative of the three metrics: people who have interacted with your site recently are the most likely to interact again.

We aggregate user recency and sign-up data to track some important metrics, including:

• 7-day growth - the number of users gained in the last seven days, divided by the total number of users.  As a start-up, we're focused on building growth, so we watch this carefully.
• 30-day retention - the percentage of users who have come back to the site in the last 30 days.  It's easier to retain an existing customer than acquire a new one, so a healthy business has high retention.

Tracking Swivel Metrics (Part 1)

I've recently seen a handful of discussions like this one calling for better business intelligence education.  At Swivel, we feel that the community can be a great resource for learning.  So, in the spirit of self-education, we'd like to share the metrics that matter most to us.

Over the next few days, we'll talk about the numbers we track.  Today, we'll start with visitors.

Visitors
We track visits to understand Swivel's potential: visitors can become users and users can become customers. We use the free tool Google Analytics to capture statistics about visitors. The key indicators we focus on include:

• Daily unique visitors - The number of daily visitors to the site. If you use Google Analytics, beware that its handling of unique visitors is a bit misleading.  Limit your unique visitors report to a single day in order to get data that is comparable day-over-day.
  
• Bounce rate - The percentage of people that leave the site after visiting just one page.  A high bounce rate means that we aren't drawing many people deeper into the site.
• Time on site - The amount of time people spend on the site.  Some of Swivel's pages are very interactive, so time on site is often a better indicator of engagement than bounce rate.

Obama v. McCain: war of the words

Following the first Presidential debate that took place Friday evening, Obama and McCain both launched ads targeting their opposing candidate's choice of words.

Early Saturday morning, Obama's campaign released an ad criticizing McCain for failing to say "middle class."  Obama managed to mention the "middle class" three times during the debate on foreign policy.

McCain's campaign released an Internet video citing occasions where Obama said that he agreed with McCain's positions.  Obama said the word "agree" five times during the debate; he also said "disagree" three times.  McCain also criticized Obama for not using the word "victory" in reference to Iraq.  McCain said the word "victory" twice and referred to the consequences of "defeat" ten times.

Here's a closer look at what the candidates were talking about using some nifty data visualizations:

Total word count:
Obama: 7,657
McCain: 7,197

Wordle clouds:

Frequently used words:

The authenticity of data

Is what you're reading online true, false, or somewhere in between?

The Internet has done wonders in terms of making information accessible to people all over the world. Today, bloggers, citizen reporters, and even members of the Swivel community are responsible for adding to the so-called data deluge.  As more and more people pump information online, should we be asking the question: how much of it is actually reliable?

To determine which blogs are more trustworthy, Technorati assigns each one an "authority number".  The more people that link to a blog, the higher the number.  Of 4,969,251 blogs, more than 75,000 have an authority number of 50 or more, which means that at least 50 other people found their blog worth linking to.

At Swivel, data that comes from our Official Sources, such as OECD or WHO, is branded with an official source badge to let our users know that the data is coming from a reliable source.  Swivel members can also link back to a data source for added credibility.  Of the 13,138 data sets on Swivel (as of 9/23/08), 9,049 have been uploaded by the Swivel community and 4,087 have been uploaded by official sources.

While both systems have their benefits and flaws, a number or a badge can only tell you so much. In the end it's really up to you to weed out what's valuable from what's rubbish.

Watching Election Data

Though we're a bunch of data nerds here at Swivel, we still talk about the same things as everybody else. Recently, the elections have captured our interest.

Perhaps unlike most people, we at Swivel like to watch the drama unfold in the data.

We monitor news and media attention with Google Trends and Nielsen's BlogPulse.

And we track the results of the changes in media attention with Gallup polls and Intrade elections futures.

Hat tip to Data Mining: Text Mining, Visualization and Social Media for the inspiration for this post.

Join the Swivel Data Team!

When you went to see The Dark Knight (if you haven't, stop reading this, and go... now), did you find yourself wondering how the film's buzz could affect the numbers at the box office? When you heard Barack Obama was going to give a speech in Berlin, were you itching to analyze the word count of the transcript? When a crime takes place at the local college campus, do you feel the urge to calculate the crime per capita in college towns across the nation? Or, when you buy milk at the grocery store, does the prevalence of osteoporosis flit across the back of your mind?

If you found yourself nodding in response to any of these questions, you should join the Swivel Data Team! Our data correspondents find compelling data, upload it to Swivel, and share their insights through featured graphs on the home page. For more information, check out the Swivel Data Team, or you can .

5 types of data visualization people

Whether you're looking at a scatter plot or a Jackson Pollack painting every person notices and appreciates different things.  Flowing Data has defined five different types of data visualization people by what they see in the data and what they take away from it.  The five types are: the technician, the analyzer, the artist, the outsider, and the light bulb.  Check out the blog post for a more detailed description of each type.

Ten people responded to the post and many defined themselves as hybrids of two or more types.  Here are the results of their responses:

Which type (or types) of data visualization person are you?

Data language survey results

If you'd like to see the results of the survey as the data come in you can check them out here:

Survey results

Please forward the survey link to other folks you know who care about data: survey.

If you have additional commentary go ahead and share your thoughts on this blog post or send me an email brian@swivel.com.