Visual Communication Lab Blog

We are happy to announce our most recent project, Historio (http://historio.researchlabs.ibm.com). Historio is an experiment in storytelling and visualization; we are interested in helping people tell stories by structuring their narrative into interactive timelines.

The first incarnation of Historio celebrates IBM’s centennial anniversary. We invite IBMers worldwide to contribute their stories to our system. Members of the public can use Historio to explore IBM’s rich legacy and share their thoughts in the comments section at the bottom of each story. Your favorite histories can be bookmarked and shared on several social networking websites.

Stories in Historio are structured as timelines and consist of time periods and discrete events that pair the author’s narrative with photos, videos, documents, and geographic information. On our history exploration page you can search for histories by keyword, location, or industry, and see the results on our circular overview visualization.

Try clicking on the timelines in the circle to see each history in more detail. You can open several histories at once to compare them.

Click the button at the bottom of each column to view the complete history, including all of the media associated with it and the comments left by readers like you.

We hope Historio will help you discover something new about IBM, and we look forward to hearing your feedback.

Thanks and we hope you enjoy it http://historio.researchlabs.ibm.com
VCL

The VCL is pleased to be a part of ‘Telling Stories With Data: The Next Chapter’, a workshop to take place at the IEEE InfoVis 2011 conference this October in Providence, Rhode Island. VCL researchers Joan DiMicco, Adam Perer and VCL intern Jessica Hullman join Karrie Karhalios of University of Illinois and Nick Diakopoulos of Rutgers University to organize an exciting day. Expect great discussions and presentations from leading designers and scholars, Sunday October as IEEE VisWeek 2011 kicks off!

This workshop is the second iteration of Telling Stories with Data. Last year’s workshop brought together dozens of visualization researchers, journalists, humanities scholars, and tool builders to talk about how data has the potential to promote increasingly sophisticated and data-literate conversations to the world at large. As the use of and interest in storytelling and visualization have only increased in the last year, this year’s continuation promises to be an even bigger success!

You can take part in the workshop by submitting a 1 to 2 page position paper by September 2. You can find more details on what to include in your submission here:

http://data-stories.com

One of the motivations behind taking a closer look at the use of data visualizations in storytelling contexts stems from the important role of communication-minded visualization: the construction of narratives that are founded upon and illustrated by data. Ideas around this topic began to be discussed at the VCL several years ago, as tools like Many Eyes were launched, and users relied on the visualizations as communication devices to be used elsewhere online, such as in blogs. When average web users encounter visualization, it is typically presented as part of a narrative, such as a blog post or news article. When a user creates or shares a visualization, it is often with the intention of sharing a story of their own. The data this user visualizes may be personal, communal, or public; regardless, the visual representation of the data is a critical component of their narrative. Designers and users of visualizations alike rely on techniques for narrating or framing presented information though many of these techniques are not captured by traditional infovis design guidance. Understanding how data and storytelling interact deepens knowledge of visualization, as it is through storytelling that people structure and share the insights they glean from visualizations.

For a great baseball-related example of visualization used in storytelling, check out the New York Times’ use of visualization in telling ‘How Mariano Rivera Dominates Hitters’:

http://www.nytimes.com/interactive/2010/06/29/magazine/rivera-pitches.html

To see more examples and find out more on workshop goals and topics for discussion, check out the website:

http://data-stories.com/

You can also like the Facebook group to get announcements as the date draws closer!: http://www.facebook.com/data.stories

Last week, I had the honor of giving a keynote talk at the ACM/IEEE Joint-Conference on Digital Libraries, in which I talked about how interactive visualizations can be used as storytelling devices and in some cases support storytelling more so than analysis. This is a theme our group has been thinking about for a year or so, beginning with our InfoVis workshop last year, Telling Stories with Data. (Read more about the workshop in our guest blog post on Flowing Data.)

This JCDL talk started with an overview of what I see as the components of a traditional story that should be present in an interaction with a visualization for it to become a story. These components are:

• Concrete: There are specifics about people, things, and/or events.
• Temporal: Aspects of narrative and chronology, sequential presentation.
• Purposeful: Agents have goals, conflicts, and resolutions; A question is posed that comes to resolution.
• A storyteller and an audience: There can be many possible storytellers to same data story.
• Emotional: This is a key point that distinguishes impartial data analysis from a story.

A basic example of a data visualization that supports stories, and actually is not meant for serious analysis, is wordle.net. The focus of my talk was on the stories that are told around the visualizations in Many Bills, SaNDVis, and Second Messenger. Check out the slides below:

Without further ado we’re excited to announce the first of our sketch projects: NYT Writes created by Irene Ros, a research developer in our lab. This sketch looks at The New York Times writers and the topics they write about. It compares the diversity of topics by authors to what a single topic is comprised of.

How does it work?

You begin by performing a search for a topic of interest. Pick a keyword you’re interested, such as “Tsunami”. This will fetch articles containing that term that were written in the last 30 days and build the visualization from them.

What am I looking at here?

There are a few things that you will see once the search is complete. First, on the left side of the screen you will see a stack of bubbles at varying sizes. Each bubble represents a term, or “facet”, that was used to describe one or more articles containing your search query. Facets get manually attached to each article by The New York Times staff. An article about “Tsunami” might be tagged as being about “Natural Disasters,” for example. The size corresponds to the relative amount of times that tag appeared comparing to all the other facets collected from all other articles in the query set. You can mouse over each bubble to see the tag name appear in the middle as well as how much it appeared relative to the other facets below the stack itself. This stack could also represent what I call a “dedicated writer” – someone who only writes about one topic for 30 days would have a similar stack to this one.

Hovering over a label also reveals its relationship to other facets through arcs. The white arcs that connect two bubbles indicate that the two facets were used at some point to describe the same article. The thicker the arc, the more times the pair was used across all articles in the query set.

What does the author list on the side mean?

As we collect articles into your query set, we also extract the authors that appear in the byline of the article. The list on the right side contains the authors we extracted alongside the number of articles they appeared on in the query set. They are sorted alphabetically. You can click an author’s name to get another facet bubble stack to appear beside your query bubble stack.

What does the author stack mean?

When you chose an author to review, we go ahead and fetch the articles they wrote in the last 30 days. We then collect the facets from those articles and build a bubble stack much in the same way we built the first one. The key here is that an author might (or might not) write about a diversity of topics and using the same visual structure we can compare how closely an author’s writing resembles the original query topic. In most cases, authors write about so many more topics than just the one we query for, showing us how versatile they have to be. The bubble sizes in the author stack are relative in size to those in the query stack. Mousing over a bubble in one stack will highlight that same facet bubble in the other stack.

When you’ve selected an author, you can see how closely they resemble the first bubble stack at the bottom of the visualization.

The similarity is computed via a standard metric called the Pearson Correlation Coefficient.

Where do I try it out live?

You can find it at http://nytwrites.thevcl.com.

Questions?

Feel free to contact me at imirene at gmail dot com.

We’ve been a part of the wonderful visualization community for years now and participated in many ways. While we each have our own way of communicating our ideas, we decided we also have thoughts we’d like to share as a research lab. So….. Hello, Blogosphere! We will be using this blog to share our thoughts about visualization and also to share updates on our current work.

And right off the bat, we’re incredibly excited to announce a new effort underway called VCL Sketches, which are short visualization and design experiments we will be periodically sharing online.

In general, the Visual Communication Lab works on large scale projects like Many Eyes or Many Bills that take months to develop and then launch, but meanwhile, we are also trying out new technologies for the sake of learning and playing with interesting datasets. We figured we should share those with you all!

So in our next blog post we will announce the first VCL Sketch. Keep an eye on this blog for future Sketches.

• 

  Joan DiMicco

• 

  Adam Perer

• 

  Jenn Thom-Santelli

• 

  Irene Ros

• 

  Yannick Assogba

• 

  Jamie Rasmussen