We tracked down Eric Fischer, a data visualization artist featured in the Atlantic, New York Times, Washington Post among others, who creates maps using geotagged Flickr uploads or Tweets. Below is his vivid tell-all about taking large datasets and making them meaningful, even beautiful.
What is Data Visualization?
Pattern recognition is what our eyes and brains have evolved to do. If all you have is numbers, you can only make connections through conscious thought and careful calculation, but if you can turn it into an image, the relationships between things immediately pop out. Data visualization gives us access to data in ways that are more tangible and meaningful.
In what ways are you telling a story through your maps?
The same map can tell lots of stories, but fundamentally what I am interested in is knowing what places people like to spend time in and what places they avoid, and what the systematic differences between those places are so that we can make more places where people like to be.
How do users react to viewing their city mapped out on Flickr?
People seem fascinated by maps of their own cities if the map confirms what they know about the places they know and lets them learn something about the areas that they don’t know so well.
How can data scientists or software programs refine “crude” data for users?
Most of the refinement that I do is to eliminate misleading clusters. Some people post lots and lots of pictures or tweets from a very small range of locations, making it look like that location is very popular when it is really just one person. Filtering out duplicates within a small radius makes the visualization much more representative of the general pattern.
Personal data without the person- are there any risks associated with population data visualization?
The most notorious abuse of population data has been by governments trying to suppress minority populations, most notably in World War II. But this is more of a result of data collection than of visualization: if a government is trying to suppress minorities, it will probably do it by the numbers instead of by mapping.
Can we see a future of social media & data visualization working together to monitor incidences of disease?
I don’t think this will work very well. Google Flu Trends works reasonably well because people using Google search are looking for solutions to their problems. Reporting or tweeting symptoms of illness would differ from the way people would normally use Twitter.
Art galleries must be knocking on your door, any plans to sell your work?
Thanks! I have been meaning to make prints for sale and am talking to some people later this week about trying to get things set up.
The Green details the movement of travelers and the Purple is a reply on twitter. The white represents the most interesting piece of data- where travelers and replies follow the same direction.