(5 December 2018) SAGE Publishing, a leading global academic publisher, today announces that it has made a major seed investment in Thusly Inc. – a technology start-up creating research tools for big data analysis of document archives. The investment is the first for SAGE as part of its SAGE Ocean initiative, and will support the launch of Thusly’s TagWorks crowd annotation software into the academic research market.
TagWorks enables social science researchers to perform complex content analysis on very large document collections at up to ten times the speed of existing methods. The collaborative tool enables researchers to deeply analyze documents by asking internet-based workers a series of detailed questions to evaluate passages of text. TagWorks is currently being used to identify and report nuanced forms of misinformation in news articles such as ‘exaggerated metaphor,’ and ‘confusion of correlation and causation.’
Thusly Inc. was co-founded by Nick Adams Ph.D, a sociologist and former research scientist at the Berkeley Institute for Data Science, and Norman Gilmore, a seasoned software developer with a long term interest in citizen science. The idea for TagWorks was born when Adams’s research posed the challenge of closely annotating nearly 10,000 news articles describing events of the Occupy movement.
“Wherever researchers have big data in the form of documents –– whether that’s government archives, newspaper data, court transcripts, meeting minutes etc. –– they have data that has, until now, been practically inaccessible,” says Nick Adams, CEO of Thusly. “TagWorks will open up new horizons for social science. Analyzing big archives used to mean training waves of research assistants, and content analysis projects would stretch on for years. Now, hundreds of internet-based analysts can work on a project with no face-to-face training, meaning a job that would have taken a decade can be completed in a matter of months. What’s more, researchers don’t have to choose between quantitative and qualitative data, they can have both. Access to such big, rich data enables us to ask and answer complex questions we couldn’t before.”
SAGE Ocean was launched earlier this year to support social scientists working with big data and new technologies. TagWorks joins its growing portfolio of products, which also includes social data science e-learning platform, SAGE Campus.
Katie Metzler, Executive Head of Methods Innovation at SAGE says, “This investment is an example of the type of innovation we want to catalyze with SAGE Ocean. Software like this needs investment and incubation to make it from an inventor’s sketchpad into the hands of social scientists. This is an important milestone for us as we build a portfolio of new products to meet the evolving needs of social science researchers working in the age of big data.”
Nick Adams continues, “We’re really grateful to SAGE for recognizing the promise of TagWorks. It takes some methodological expertise to understand the deep value of this tool, and a real commitment to social science to provide the kind of support SAGE Ocean has offered. The incubation team has coached us on the development of our business plan, pricing model, marketing materials, and business operations. We really couldn’t hope for a better strategic partner to ensure that this powerful tool is enabling the next generation of big data research.”
TagWorks’ first customers include the Public Editor project from Goodly Labs, a California-based non-profit evaluating thousands of news articles for social good projects (SAGE Publishing announced a sponsorship of Public Editor earlier this year) and the History Lab at Columbia University, which is using TagWorks with National Science Foundation support to annotate and curate U.S. State Department cables. TagWorks will also provide ideal human-labeled “training” data for research projects developing artificial intelligence through supervised machine learning.
Matt Connelly Ph.D., the Director of Columbia’s History Lab, says his team feels fortunate to be one of TagWorks’ first customers. “We have so many documents we’re curating for the broader research community and we just couldn’t find another tool that would allow us to work at the big data scale. Automated techniques produce too many errors. But old-style annotation tools aren’t designed to work for such large archives. We had thought a few years back that someone should create a crowd approach for annotating large archives. With TagWorks, we’re glad it’s finally here.”
The full announcement is here.