Rob Diana recently did a post suggesting Facebook, rather than Twitter, was the real goldmine for data. I, as I’ve inferred before on LouisGray.com, wholeheartedly agree with that notion, and in fact, Facebook is already showing the potential for this with their Lexicon product. While not yet as public as Twitter search, right now anyone can see top trending keywords and topics, very similar to Twitter search, via the URL http://facebook.com/lexicon. Better yet, Facebook has revealed where they are going with it via their new version of the trends tool at http://facebook.com/lexicon/new.
From the new Lexicon you can pick any trending topic, see the number of users talking about those topics on a given timeline. Facebook goes further though, providing demographic data. For instance, for the term, “baseball”, you can find out how many females vs. males are talking about baseball. Under a Demographics category, you can also break it down by age, or country.
In addition to demographical information, you can pull out various statistics about what people are saying about a particular topic. Under the “Associations” category, you can get a graphical breakdown of what the most popular word associations with the topic are. Popular associations for the word, “baseball” were “high school”, “college”, “coach”, to name a few.
I found this one quite interesting. Facebook has found some way to determine the sentiment of users when mentioning various search terms. For instance, for “baseball”, you can determine the percentage of users speaking positively of baseball, and the percentage of users speaking negatively. They go further though, allowing you to compare the sentiment of one search term over another.
Facebook will also do some biographical checking for you as well. It looks at each search term, and will determine what other common keywords are in users’ profiles related to those search terms, and what words are most popular when users are mentioning the term. For “baseball”, users tended to have words such as “sports”, “music”, “football”, and “reading”, revealing a little information about those fanatical about baseball.
Lastly, users can see how a particular search term fares geographically. “baseball” is very popular in California. Facebook provides a map, color-coding based on percent of users talking about the particular term. I can’t help but wonder if all search results are biased towards California however since that is one of the most populous states in terms of Facebook use.
The Potential for Facebook vs. Twitter
I think Facebook has shown through this Lexicon that they have the potential to be much more useful than Twitter in terms of search and data mining potential. Because Facebook has more detailed profile data, and a significantly larger user base to read from, the potentials for useful data are so much greater, and are already proving so via this Lexicon. If Facebook starts to provide APIs around this search data, along with the publicly available user status updates and profile data, they will be a very serious force to reckon with, that I think, regardless of the mass funding Twitter has, will be extremely tough to compete with.
If you’re currently writing apps for Twitter, you should sincerely consider starting to learn Facebook.