When two Boston-area organizations rolled out an interactive data visualization website last month, it represented one of the largest public uses yet for the open-source project Weave -- and more are on the way.
Three years in development so far and still in beta, Weave is designed so government agencies, non-profits and corporate users can offer the public an easy-to-use platform for examining information. Want to see the relationship between low household incomes and student reading scores
in eastern Mass.? How housing and transportation costs compare with income? Or maybe how obesity rates have changed over time? Load some data to generate a table, scatter plot and map.
In addition to viewing data, mousing over various entries lets you highlight items on multiple visualizations at once: map, map legend, bar chart and scatter plot, for example. Users can also add visualization elements or change data sets, as well as right-click to look up related information on the Web.
The benefits of Weave's interactivity go beyond the visual appeal of selecting an area on a chart and seeing matches highlighted on a map, said consultant James Farnam, project coordinator for the Connecticut Data Collaborative and an early Weave backer. "You're creating subsets of data on the fly," he said. With a single click on a scatter plot, "you can recalculate regression lines and relationships that you're testing."
Users are already working on a set of quality tests within Weave to help find data errors visually, he said.
Data visualization tools have long been in the hands of the technically savvy, but Weave aims to help organizations democratize them, creating what project head Georges G. Grinstein calls a Wikipedia of data -- a way for anyone interested in a topic to explore and analyze information about it, instead of leaving the task solely to computer and data specialists.
"Now [you're] engaging the public in a dialog with the data," said Grinstein, director of the University of Massachusetts at Lowell's Institute for Visualization and Perception Research. "That's why Weave is open source and free" -- even though it contains some university-patented technology (the institution agreed to allow it in the software).
Weave is "ridiculously powerful," said Holly St. Clair, data services director at the Metropolitan Area Planning Council. The MAPC is using Weave in its MetroBoston DataCommon site, created jointly with the Boston Foundation's Boston Indicators Project. "The power that we see and the versatility is amazing." In fact, one of the challenges of implementing Weave was how to narrow down its offerings so that end users wouldn't be overwhelmed with options, she said.
Another issue is basic to a lot of early-stage open-source software: limited formal training options for staff compared to more established commercial products. However, she believes that will change as Weave becomes more widely adopted.
There are about 25 organizations that have been using Weave and giving feedback, including 10 since the project's beginning. "Each one had a whole set of different requirements," Grinstein said. "The technology is so rich because of the first 10 users.... We're driven by requests."
Farnam called the interaction between consortium members and UMass students and faculty "pretty remarkable," with features being regularly added and updated during an agile development process as the software evolved through version 1.0.
About 25 to 30 students have worked on the project in its first three years at UMass-Lowell, in partnership with the Open Indicators Consortium, a group of early users and supporters of the project. Grinstein expects work will continue for another three years at UMass-Lowell, and involve whatever additions the open-source community wants to contribute. The project was built using Adobe Flex and ActionScript.
Several more powerful features have already been architected and are just awaiting user-interface design, including collaboration and session-capture expected this summer.
Collaboration will allow people in multiple locations to work on a visualization together in real time, without needing a screen-sharing application such as WebEx, Grinstein said.
Session-capture will let users record every step they do in making a visualization, so they can re-create the process for another visualization or share their steps with other users. Once privacy issues are worked out, Grinstein said, such session captures could also be used by researchers to better understand how people interact with data -- and even offer suggested next steps to new users if they get stuck.
Weave is still somewhat difficult to install, Grinstein admitted, but plans call for a lighter one-click installer by summer as well.
Also on the way: so-called "infomaps," one of the patented technologies within Weave, that can tie a mapping visualization to a collection of documents. Even if a document isn't geocoded but just mentions, say, "Andover, Massachusetts," Grinstein said, that document would be retrieved if a user clicked on Andover on the Weave-created map. It is, Grinstein said, like Google Maps tied to a body of documents -- while also offering multi-visualization interactivity.
St. Clair said she's become so used to working in Weave that she finds herself getting frustrated in Excel because she can't simply mouse over numbers and see matching data highlighted on a nearby chart.
"You start not being satisfied with filtering data in spreadsheets," Farnam agreed.
Interaction, St. Clair added, "starts to morph the way you think about data."