Center'd Gets A Facelift, Introduces Semantic Analysis For Smarter Local Activity Guide

Center’d, a local activity guide headed by former Yahoo Local GM Jennifer Dulski, is getting a major upgrade today. Alongside a completely revamped homepage, the site is launching a reworked search engine that it says should outperform the keyword searches found on most other local sites.

Center’d has compiled a database of around 1 million entries for various activities, each of which is categorized into a number of intent-based classifications. To do this, the site has spidered through the web analyzing ‘conversations’ taking place around each entry, taking context into account to determine if a review or comment is positive or negative. It then maps out the results in bar graphs, as seen below. Dulski says that this kind of semantic analysis is better than standard keyword search, and it helps eliminate inaccurate matches – for example it would prevent a review that said “this place is not for kids” from appearing under a query for restaurants “for kids”.

Using this database, the site can also generate city guides for users with a variety of different criteria (for example, you could generate a guide for San Francisco with romance in mind, or you could create one that would take you through the city on the cheap). The site is launching with support for twelve cities intitially, with plans to ramp up to more in the near future. Dulski says that these guides are mostly-automated (which will help it scale), though there is some editor control involved.

Center’d emerged about a year ago from the ashes of Fatdoor, a social network for neighbors. Until now its primary focus has been to serve as a local search engine and event planning site, and now it’s adding a new goal to that list: helping people figure out what to do with their day. Dulski says that many people have been coming to the site to find something to do, without anything in particular in mind. As with the city guides, users can select from a variety of criteria like ‘cheap’ or ‘for kids’, and ask the site to generate a list of possible activities.

In the next few weeks Center’d will also be deploying its suggestion engine to the iPhone, with a mobile application that will allow users to generate a day-long itinerary based on the amount of money they’re able to spend and the type of activities they’d like to persue.