Raul Valdes-Perez has long been fascinated by the search.
The former co-founder of Vivisimo, acquired by IBM in 2012, Valdes-Perez has always enjoyed the thrill of discovering knowledge buried deep in databases, then taking that knowledge and presenting it articulately.
Now he is at it again, taking the search to the next level with a new startup called OnlyBoth. The platform, developed with collaborator Andre Lessa, digs and retrieves insightful information on specific topics and writes it up in almost perfect English sentences. This two-fer aspect of the process—especially the language translation—makes it wholly unique in the world of data mining results, he says.
“Some (of the insights) are a head-scratcher and some are insightful,” he admits, “but isn’t that true of human intelligence? The general idea is a technology that generates insightful information in cases where knowing distinctive things matter.”
The inaugural version of OnlyBoth delves into the realm of higher education—rankings, geography, research dollars and other fun collegiate facts—to reveal insights beyond what the human mind might learn. Think Trivial Pursuit on big data steroids.
The process is simple and fun. Type in the name of any 4-year university or college into the search box and within a split-second, a list of insights is generated. Some are interesting. Others are indeed head-scratching.
A few examples:
“Carnegie Mellon is the only college whose top Masters (sic) major is information technology.”
“Harvard has the highest 75th percentile SAT reading score (800) among all 3,122 colleges.”
Or Valdes-Perez personal favorite—”UCSD has the most members of the National Academy of Sciences (67) of the 1,049 colleges with a 25th percentile SAT reading score of 520 or less.” (He likes that it reveals a lack of verbal strength at UCSD.)
OnlyBoth’s artificial intelligence technology was first developed when Valdez-Perez was at CMU in the late 1990s with the help of a National Science Foundation grant. The idea was shelved when he shifted his focus to Vivisimo, he says. Most of the data is pulled from the federal government.
While these predictions may eventually pave the way to a better understanding of consumer behavior and other whatnot, OnlyBoth will remain a fun and free application for the public for now. Data mining and big data are about making predictions, he says, alluding to the power behind the data.
The value proposition exists, he agrees. Especially in OnlyBoth findings like this: young men who buy diapers in a supermarket on Fridays often also buy beer. Where it leads remains to be seen.
An application for sports and teams may be next.