You’ve probably heard about this whole “big data” thing. So you have lots of data collected from lots of users sitting on lots of servers. Cool.
Your ability to do anything really interesting with that data is partly limited by the power of the computer in front of you, which is fine at a small scale.
But when you consider the fact that Facebook employees sift one million gigabyes of data per day in search of ways to make the social network better, suddenly that laptop looks a little bit more like a calculator watch.
Sense, a startup that launched today after raising a seed round from the likes of Facebook co-founder Andrew McCollum and former Microsoft chief economist Susan Athey, thinks it has a better way: Giving more power to the maths nerds.
At first blush, Sense’s promise may sound like a lot of companies that promise to help you make better use of data. Giants like Microsoft and startups like Birst and Chartio offer so-called “business intelligence” tools, for instance.
But those tools are primarily designed to offer ways for businesspeople to play with data in a super-simplified way that mathemeticians would scoff at. The inflexible maths in many of these tools makes it hard for anybody who really knows what they’re doing to get more than basic answers.
This is where Sense comes in. It’s designed to offer data scientists — like Zajonc and co-founder Anand Patil, who are both PhD-holding statisticians in addition to coders — a set of tools to do the kinds of maths they want to do, and the power to do it. This is for medical researchers, financial services companies, big web services, or anybody who works with large amounts of information and needs to ask big, complicated questions.
Sense believes the tools that data scientists have had until now are too slow, too hard to use, and make it hard to share those results.
Zajonc says Sense solves a lot of those problems by providing every data scientist with access to a supercomputer, thanks to the fact that the platform runs on a cloud hosted in massive data centres elsewhere. And with a slick browser interface and integrated collaboration tools, it makes it easier to those data scientists to read and share their analyses.
Under the hood, Sense users can write scripts in the very popular R, Python, and Spark languages, plus share their code (and results) with other data scientists on their same team.
Big data is big business, with companies like Cloudera and Hortonworks raising plenty of money to build the kind of so-called “data warehouses” that hold the data that Sense works with. Sense just wants to help mathematicians do more with it.