Summary:

Startup Prior Knowledge opened up its public beta to its database API on Monday, so it can solve the problems of developers who want to play with data, but who’d rather avoid all that pesky math. Prior Knowledge has raised $1.4 million to achieve its goals.

Eric Jonas, CEO and co-founder of Prior Knowledge

Eric Jonas, CEO and co-founder of Prior Knowledge

Startup Prior Knowledge opened up the public beta to its database API on Monday so it can solve the problems of developers who want to play with data, but who would rather avoid all that pesky math.

The need for data analysis often starts with a hunch. But somewhere between trying to figure out if the parking meters near a local police hangout are generally ticketed faster than others, you realize that aside from the data on where parking tickets were given and how often, you may need more info and you still aren’t sure what math to perform to prove a relationship. Generally, that’s where most people give up.

But Prior Knowledge wants to enable app developers who aren’t data geniuses to keep playing their hunches by offering up a service that helps them figure out what they need to come up with a statistically relevant answer. Like other startups including ClearStory, Datahero and Platfora, the 10 employees at Prior Knowledge decided to step into the gap created by the vast quantities of data and the dearth of people who have the stats knowledge to do anything smart with it.

The San Francisco-based startup, which was formed in August 2011, is the brainchild of former MIT graduates who saw the rising tide of data excitement and the looming shortage of “data scientists.” The startup raised $1.4 million in February from the Founders Fund, but has been building its product, Veritable, which is hosted on the Amazon Web Services  cloud, since October. 

“Our goal is to use the state of the art in machine learning to create math that helps people deal with data uncertainty,” said Eric Jonas, the co-founder and CEO of Prior Knowledge. In layman’s terms, it means he’s built a really smart predictive database that can look at the problems users want to solve and do the right math to solve them. Because most users aren’t quant jocks, Veritable also looks at data users put into the service and, like fellow machine learning startup BigML, points out interesting correlations they may not have seen.

A visual example of the mathematical grouping that Prior Knowledge’s database does behind the scenes.

So what exactly does Jonas want to do for devs with Veritable, which he says is unique because it can consider the trillions of ways every variable can interact with every other variable to generate predictive models?

“When it comes to data, the math gets hard really quickly and most developers — your average hacker — [are] trying to build or craft an app and [don't] know how or when to use a linear regression,” Jonas says. “That’s like learning how Unix works just to get to the ‘Hello World’ phrase.” So, for now, Prior Knowledge does that for developers, charging them by how much data they submit and by the answers they request.

Jonas says the ideal user is a developer who knows Ruby or Python and has an inkling of what they want to build. So maybe it’s app that uses data from the Center for Disease Control to determine if someone has the flu, or maybe it’s an app to help boost the success of recommendations on an e-commerce site. Maybe it’s an online store that wants to sell better-fitting bras. The developer knows about the user experience, but maybe they don’t know as much about statistics and math.

A visual example of the mathematical grouping that Prior Knowledge’s database does behind the scenes.

The database can look at the data people upload and determine what data is actually most relevant for a query and pinpoint interesting correlations between different data sets. If someone uploads a person’s shopping history and Veritable shows that age is the best predictor of the type of shoe someone will buy, he can adapt his recommendation engine accordingly.

Jonas sums it up as giving developers the ability to “predict, explain and group.” So your apps might offer people better predictions (don’t neglect to feed the meter at this spot), explain behavior (people with these symptoms have the flu) and group data (shoppers in this cohort buy Product X).

At least two application developers have been using Veritable during the private beta period, Jonas said, one building a tool for financial risk assessment for hedge funds and the other associated with healthcare. Whether it will become the go-to platform more startups needing data analysis platforms remains to be seen, but the service is now public beta, so I guess we’ll soon find out if it has mass appeal.

You’re subscribed! If you like, you can update your settings

Comments have been disabled for this post