Yo, taxi! CabSense finds rides — and a way to crunch big data

Why can’t you ever find a cab when you really need one? This is even true in New York City, famous for its seemingly endless streams of yellow cabs gushing up and down the avenues. If only there were a way of knowing the best corner for getting a taxi ride. If you ask Sense Networks, there is: a new iPhone app called CabSense.

Sense Networks is best known in the Bay Area for CitySense, an iPhone application for gauging nightlife activity in San Francisco. CabSense, which the company just released two weeks ago, shows on a map the best street corners for getting a cab in New York, taking into account the time and day. With CabSense, the user can see “hot” corners for getting a cab in the vicinity right now, or plan ahead and see what would be the best location for hailing a taxi, say, the next day to get to the airport.

[aditude-amp id="flyingcarpet" targeting='{"env":"staging","page_type":"article","post_id":175868,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,mobile,","session":"C"}']

“Right now” doesn’t mean that CabSense will tell you exactly where the cabs are driving right at that moment, though. It does not track specific taxis in real time. But its taxicab predictions are based on some heavy-duty data crunching, said Blake Shaw, a project manager at Sense Networks, who worked closely on CabSense.

“CabSense gives users the historical average of customer pick-ups and drop-offs at different location,” Shaw explained. “Corners are given a star rating which tells you the probability of available taxis at a given time.”

He told me this on Monday at 6:10 p.m., when the corner of Broadway and 116th Street, where we conducted the interview, garnered three and a half stars.

CabSense also has a radar view (pictured here), which is a compass showing the direction and distance with a yellow arrow to a street corner with the most activity.

Getting to the historical average requires the algorithms behind CabSense to chew through massive amounts of data to calculate the likelihood of an available free taxi. So far, CabSense is working on the data from 90 million trips, and the information is continuously accumulated through regular data dumps Sense Networks gets from taxi companies and mobile phones.

The 3.5 stars for our corner of Broadway and 116th mean 20 to 30 customer pickups per hour, on average. During our conversation, the number seemed to be in the right ballpark as free taxis whizzed by at a steady rate. But one might argue that with 12,000 taxis in the city and 500,000 daily cab rides, New York is a pretty easy place to get a ride (unlike, say, San Francisco taxis, “of which there are five,” as the comedian Eddie Izzard once said).

“You would be right to think that for Manhattan,” Shaw said. “But, this may not be true for the other boroughs like Brooklyn or Staten Island. And sometimes the oddest places in Manhattan get more traffic than the Grand Central Terminal or Penn Station, which get a lot of heat during mornings. For instance, the corner of Essex and Rivington [on the Lower East Side] around 11 p.m. on a Friday gets 66.9 pickups per hour, which means there are a lot of popular bars around that area.”

CabSense and CitySense are only promotional tools for Sense Networks. The apps are free and fun to use, and since the launch at the end of March, 7,500 users have installed CabSense. However, their purpose is to demonstrate what the company can do with a large mass of information in terms of prediction and machine learning.

[aditude-amp id="medium1" targeting='{"env":"staging","page_type":"article","post_id":175868,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,mobile,","session":"C"}']

“There’s no money in CabSense,” Shaw admitted. “Our products are the algorithms. Our clients can be carriers or app developers or advertisers, anyone who needs tools for analyzing data such as location intelligently. Information about location gives so much context, and we can leverage that context to provide better predictions for future activity. What business wouldn’t be interested in knowing where I’m likely to go next?”

One example would be the aforementioned mobile carriers, who might be interested in figuring out churn rates, or the number of subscribers changing carriers.

“Let’s say you go to a lot of bars in an expensive neighborhood. That could mean you’re young and you have some cash to spend. Chances are, you could also be interested in a cool phone, and that could be relevant information for a carrier with such a phone,” said Shaw.

This doesn’t mean that CabSense and CitySense are one-off products for Sense Networks. The company is releasing an application programming interface, and 2.0 versions are in the works for both apps. (CitySense will debut on the Android platform, too.) CabSense will have real-time weather data, which not only affects the number of people wanting to get in a cab but the routes taxi drivers choose. All this leads to better predictions.

[aditude-amp id="medium2" targeting='{"env":"staging","page_type":"article","post_id":175868,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,mobile,","session":"C"}']

It’s fairly easy to predict that Sense Networks’ technology would complement some of the location-based services out there, such as the check-in games Foursquare and Gowalla or social networks like Facebook or Twitter. According to Shaw, the company would love to integrate their service with as many sources of information as possible, though he is not saying anything specific about the company’s future plans.

As far as competition in this type of analytics goes, Sense Networks is operating in fairly uncrowded space. That being said, there are giants like Google who are interested in this area, and have all the necessary resources at their disposal.

Three years old and employing twelve people, Sense Networks closed a Series B funding round last June to the tune of $6.4 million. The company’s investors include Intel Capital, Javelin Venture Partners and a number of hedge fund and angel investors.

[Photos: JP Manninen]

[aditude-amp id="medium3" targeting='{"env":"staging","page_type":"article","post_id":175868,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,mobile,","session":"C"}']

[This story is part of a weekly series on location-based services, written by VentureBeat’s JP Manninen. If you have an idea for a story you would like to see in this series, drop a line to jp@venturebeat.com]