BooRah - Inside the Buzz

We’re staging a new look for BooRah here. Over the past few months, we’ve received a lot of feedback and we’ve tried iterate on some of it. The biggest area of focus was to reduce the clutter of links and text and organize the page flow a bit better.

Do the web pages look better now?  Did we reduce some of the link and text clutter from our current version of restaurant search website here.

You’ll also notice a new feature on our city profile pages and restaurant detail pages to showcase some of our semantic meta data. Each restaurant detail page now features “Top Sentiments” from all the reviewers across all the sources we’ve aggregated and you can instantly get a gist of what’s popular at that restaurant. The city profile page now shows, “best-of” categories for you to explore various dining options. E.g Mountain View city page offers recommendations such as “Best Weekend Lunch ” at Cafe Baklava, “Best parties” at Xanh Restaurant and “Best Lunch Buffet” at Amber India.

Take our new website for a spin and let us know what you think.  Please leave feedback in the comments section here.

Yahoo, this morning announced the winners of it’s SearchMonkey contest. And, we are glad to have won the first prize in the InfoBar category. Special congratulations go out to our team member, Shrisha Radhakrishna, who spearheaded the integration effort.

SearchMonkey is the first of many initiatives for Yahoo’s Open Strategy (See ReadWriteWeb story). Earlier this week, Yahoo also announced a program to open up their existing API’s for unlimited use with no restrictions on branding/linking. We’re excited to make use of Yahoo’s open platform, with our semantic technology, to enhance the user experience and relevance for our customers. Kudos to Yahoo for taking such a bold step in the tumultuous world of Search.

Now for the details of our InfoBar implementation:

InfoBar is an application within SearchMonkey that allows additional information to be shown along with the search result. Below is a search result for “best restaurants in san jose” that has been enhanced by InfoBar (NOTE: You’ll need to be logged in to Yahoo and add the BooRah application in your profile to see these results. See post for instructions).

By expanding the search result, you can see the enhanced result within the same page.

You’ll see the top 3 restaurants and their overall ratings along with a link to start exploring the site, providing a preview of what to expect when a user lands on the site. For this result example, “Falafel drive Inn” received a 93% positive rating based on 353 reviews, “Maggiano’s” scored 89% out of 324 reviews and “Original Joe’s” scored 82% based on 290 reviews. This information was all derived from the sentiment in the written text of the reviews using BooRah’s patent-pending NLP engine.

Here’s another example — a search for “oola in san francisco“. The specific restaurant search is enhanced by displaying the most relevant attribute “tags” that were extracted from the user reviews:

In this example, “oysters”, “chicken ravioli”, “bartenders”, “fish dishes”, etc., were highly raved about in the user reviews. We also provide drilldown on ratings for food, service and ambiance to aid user evaluation before leaving the search results page.

We hope you’ll add BooRah InfoBar application to your profile and enhance your search results and relevance next time you search on Yahoo!

Today, we’re announcing the launch of our restaurant guide in Washington-Baltimore, Philadelphia, Las Vegas, Phoenix and Miami metropolitan areas. This brings the total number of metro areas covered to 15 and we are working on introducing our service to other parts of the country as well. Here are quick links for a couple these metro areas:

Restaurants in Washington DC - Baltimore Area
Restaurants in Philadelphia

In launching Las Vegas Restaurants, we encountered some interesting technical challenges in our crawling algorithm. It turns out, that most of the restaurants (atleast the popular ones) are located inside the hotels/casinos and do not have proper address and contact information in the yellow page listings. We had to tweak our entity-mapper to incorporate additional fuzzy heuristics. This will turn turn out to be very valuable in aggregating content for businesses that only have a PO Box or 1-800 telephone numbers as we expand to other categories.

We have also reached some significant milestones on the content side, a demonstration of the scalability of our semantic crawler and sentiment extraction platform. I’ll share additional details with some technical analysis today/tomorrow.