Wednesday, May 30, 2012

New Features This Week - Map and User Photos!

We are continuing to add features about twice a week.  Last week, we expanded the number of games we are covering and included all of the MLB, NHL and NBA games.  So far the system has handled the extra amount of data just fine, and we plan on covering most, if not all games, for the foreseeable future.

Within each game, we have two new tabs that rolled out this week.  The first is a 'map' tab that allows you to see the Twitter conversation of individuals within a 50 mile radius of any location that you choose.  The location defaults to the stadium of the home team, but can be customized to a location of you choice by clicking the map or typing a location in the search box.

The other feature is one that I really like.  It allows you to see user images from the game.  Here's a few from yesterday's games that we enjoyed:


Wednesday, May 23, 2012

Congratz to Kratz!

Congratulations are in order for my friend Erik Kratz on his first Major League Homerun last night in the Phillies 5-2 loss to the Nationals.  Erik has moved back and forth between Lehigh Valley (AAA) and the Phillies over the past 12 months, and made the most of his opportunity this time around with a long ball in the 8th. 

If you look at our Fan Sentiment Chart for that game, you'll see a nice little positive spike in Philadelphia fan sentiment at 7:23 (MST), which corresponds with Erik's HR.  Hopefully we'll see many more of these in the weeks and months to come!

Friday, May 18, 2012

Overlaying the Score

Our charts are even more relevant now as we've included a beta version of a new overlay.  Now as you look at the "Number of Tweets" or "Fan Sentiment" reports, you can check the "Game Score" overlay checkbox and have the score put on top of the data.  This is useful as it helps to explain the peaks and valleys within the reports.

Give it a try: Cubs vs Cardinals

Monday, May 14, 2012

Natural Language Processing

Natural Language Processing or NLP is the process of pulling out meaningful information from human text.  At Fanchirp, we have designed a proprietary NLP system that is specifically tailored for social media and for sports.  Our goal is to look at thousands of Tweets and Posts for each fan base within a game, and summarize the attitudes and sentiment.

Currently, we look through each post to determine if it is displaying any of the following: frustration, encouragement, celebration.  We also look to see if the post pertains to the officiating, if it contains profanity, if the post has a question or questions in it and the number of exclamation points.  We use all of this data in various algorithms to calculate the fan ratings that are displayed in our charts.

Here are some examples:

"Let's Go #Lakers" - This is giving the team encouragement.
"Hell yeah... #Lakers win again!" - This is a celebratory post.
"Kobe, quit taking terrible shots! #Lakers" - This is a frustrated post
"That's the 3rd bad call from that ref tonight on the #Lakers" - Post pertaining to the officiating

Friday, May 11, 2012

We are launched!

Our website has officially launched and is open to the public!  We are still in a BETA phase and are excited to have a long list of additional features that we will be adding in the near future.  There are probably still a couple of bugs that need to be corrected, but overall, everything seems to be in good shape.  Give our site a visit by going to http://www.fanchirp.com.

For now, this is how the site works:  We usually feature between 2 and 5 games a day from the MLB, NBA, NFL, NHL and/or NCAA football and basketball. With those games, we pull tweets about the game every few minutes and analyze the results with our proprietary natural language processing software.  The results are summed up into charts that show the number of tweets for each team, the sentiment of each fan base, and several other charts.  The data is aggregated into 3 minute intervals and displayed via Google's Visualization API. 

The next feature is the Conversation tab, which gives you a real-time view into the discussion about the game.  You can Tweet, Retweet or Reply to other fans directly from the application.  We also encourage users to vote for their favorite Tweets throughout the game, and the top tweets are posted the next day.


We are very excited about this launch and can't wait to roll out additional features in the near future!