“Functional visualizations are more than innovative statistical analyses and computational algorithms. They must make sense to the user and require a visual language system that uses colour, shape, line, hierarchy and composition to communicate clearly and appropriately, much like the alphabetic and character-based languages used worldwide between humans.”
(I will be presenting on visualization at an upcoming (free) elluminate online conference: Patterns and Sensemaking: Information Visualization…this post is a starting point of presentation…if you’re interested, feel free to register).
Lately, I’ve been somewhat absorbed by the value of data visualization. In recent presentations, I’ve described technology as performing a “grunt cognition” role in our efforts to make sense of complex and changing information landscapes. Consider a flickr tag cloud. The tag cloud is a visualization of the aggregated activities of many flickr users. Visualization, performed by technology, does the grunt work of creating patterns which we are then able to analyze, allowing us to move more quickly to meaning making. Visualization will play an increasingly central role in helping us to cope with information growth. A picture is worth more than a thousand words – it’s worth a thousand data sets and tags.
Visualization makes data and the attendant conceptual entities accessible to a larger population than pure data. Spend a few minutes perusing visual complexity. The non-text expression of ideas and concepts opens many doors that are closed to how we cognitively process language. Images – pictures and movies – can provide a moving and emotional message in a manner entirely different from text.
Companies like CNET have been using a “big picture” visual approach to demonstrating connections and relationships between stories. A simple image of the story or concept proximity assists in creating context and revealing relating factors. Kartoo uses visual representation of a search term to display related sites or concepts.
I’ve been playing around with IBM’s ManyEyes, a tool that allows people to upload data sets (including documents, which are then displayed as tag clouds, revealing key concepts), and then perform comparisons with data sets others have uploaded. Basically, it’s a participatory approach to data comparison.
Last week, I spoke with Martin Wattenberg of ManyEyes. Below is a rough summary of our discussion:
What is visualization?
It’s a sense making tool. It’s not about total amount of information. It’s much more interesting to reflect on accessibility. I now have access to much more. We now have access to a bunch of data bases – we don’t have time to read all the data points, visualization becomes a sense making tool.
The new feature is that data visualization is a communication tool. We are trying to give people the tools to do the complicated things.
What is your goal with Many Eyes?
ManyEyes is a research project, and a platform for future research projects. People treat visualization currently as an analytical tool. We are starting to get the sense that a lot of the value comes from people collaborating. Discovery happens when people work collectively. Sometimes, with visualization – it’s about communication, not discovery. This technology has existed since the ‘80s…but the web has made it much more popular.
What are people doing?
People are moving away from analysis to making sense. They want to know “what’s going on”, or use it for political reasons to make a point (i.e. salary comparisons), activism, visualizing illness, disease. They are trying to understand what it [the data source] means.
Some people use it for creative and artistic purposes. And more generally it’s being used for communication. The more people can play with data, the more they will understand and discover.
What does this say about how our culture views and approaches information?
It comes back to the availability of data basis…before the web, we had electronic databases, but you had to pay a huge amount to access them. Today, everyone has access. Newspapers, government are all making data available. Data is now part of the cultural discourse, as much as image and text.
A raw table of numbers is excellent. But if you’re looking at it for 5 hours it’s difficult. Visualization helps understand data.
Does the average person need a new set of “data” skills?
The most important thing is a good skeptical attitude: data is wrongly interpreted in two ways: 1) look at stats and believe it’s true 2) reflectively assume that because they’ve found a confound, they assume meaningless. All data sets are problematic. Even though there’s a lot of noise there are patterns that can assist in making sense.
Compare it to wikipedia. We seem to think text is to be perfect (Britannica) or poor (a simple blog post). Be skeptical and see that it is a mix. Problems arise from trusted data that is poor. Dealing with data is a very active process, not like watching TV. Visualization is inherently interactive. The viewer is part of the system.
What about future features on the site?
We have a few new features: 1) let the community organize itself into subgroups – it will let the site grow. 2) Data quality – our general approach is to let the community police itself. No point in trying to certify something as perfect or error-free. Sometimes data is systemically unreliable – i.e. political views – they distort data that serves their agenda. It’s important that people can explain what the source
What are your future goals for the site?
There is a pure commercial side: as you can imagine, in a big company people are looking at and trying to understand spread sheets. Right now it doesn’t work well. It’s very annoying – emailing, sharing, commenting, . If you had a system for a centralized place with access control, we can talk about the data and make decisions. Take the painful process (spreadsheets), and make people smarter through visualization.
Let’s imagine a year from now, budget uploaded in visual form, people can explore problems, run scenarios and see clear visualizations – communicate with data in a manner that makes sense. People have “fact checked” in previous elections, but it doesn’t register due complexity. If something is too complex, people won’t understand it. Simple visualizations of complex concepts changes discourse.
It’s also about the rhetoric of data. I’m not sure how well persuasion happens through audio or text. Persuasion by data-graphics can make an impact. We are giving people a tool to bring computational power of bringing visualization to the people.
We want to get to connective analysis – be smarter about these things, persuade. A discovery that doesn’t get known is one that doesn’t get communicated or understood. Rhetoric, persuasion, and communication are so important.
Teemu lists some useful resources on visualization as well, though I personally don’t believe that it is data, not knowledge, that is visualized with the tools in the post.