The New Facebook Lexicon
Facebook has just released their new Lexicon tool, which allows you too look at statistics based on what Facebook users are writing on profile walls.
Currently the tool is in a test version that only has about 20 different topics to see data for, though once it is officially rolled out you will be able to search for data with any keyword.
Fortunately for those of us in the political realm, many of the provided keywords are political: McCain, Obama, Biden, Palin, Democrats, Republicans, and politics.
Here are some of the main features of the new Lexicon.
Dashboard
The dashboard allows you to look at the number and percentages of posters and posts. Above is the dashboard for Obama. You can see that on the bottom it breaks down the posters into age and gender demographics. Unfortunately this data is raw and not weighted, so the data seems to always resemble Facebook's demographics as a whole.
Demographics
The demographics tab will allow you to see graphs of the age and gender data. Once again the fact that you can't weigh the data diminishes a lot of its potential utility.
Associations
The associations tab shows you what other terms people use in the same wall post as the keyword. This is represented graphically using circles on an XY plot. The size of the circle represents the number of posts that included the associated keyword. The X-axis indicates whether more males or females use the two keywords together more frequently, with male being to the right and female to the left. The Y-axis shows the average age of the user making the association. I think this feature may become one of the more useful of the set, but the unweighted age data affects this as well.
Sentiment
The sentiment tab shows how often the keyword is in positive or negative contexts. The algorithm/dictionary they are using for this isn't public, so I'm not sure on the specifics of how they determine the score. It is important to note that there are a lot of false positives and negatives using these algorithms. They can't determine sarcasm, so if I posted "Yeah, McCain is exactly what we need to fix the economy" sarcastically on a friend's wall, it would probably show up as positive sentiment. That being said, you are still able to get a good general idea of sentiment by looking at the trend lines. In the screen shot above you can see the compared sentiment scores for Obama and McCain over the last 14 days.
Pulse
The pulse tab shows keywords that frequently occur in the profiles of users who mentioned the topic. You can see raw numbers for interests, music, movies, TV, and books. This is definitely more useful for commercial marketers than political people, and once again raw data means that something like Harry Potter is going to be at the top of all the book lists because of its ubiquity.
Map
The map tab shows where people are talking about the selected topic(s) and allows you to compare relative popularity. Each state is colored based on how many times the topic was mentioned relative to the other topic. In the screen shot above I compared Obama and McCain. In this case Obama is mentioned more than McCain in every state. If a state mentioned McCain more than Obama it would be a shade of red.
Final Thoughts
The Facebook Lexicon might be useful for some national campaigns in the future once it opens up to all keywords, but there are some things holding it back.
As I mentioned a bunch of times before, the fact that you can't weigh data is a big weakness. What I would like to see would be to compare the data relationship to the averages of Facebook users in the United States. For example, if 50% of Facebook users have Harry Potter as a favorite book on their profile, but 58% of posters of a keyword have it listed, show the data as +8%. Conversely if only 46% listed it, show -4%. This would allow you to get a better picture of the data relationship than the raw data does.
Another thing to keep in mind is that the Lexicon only uses data from wall posts. This does not take into account status updates. I am not yet sure whether or not the final version will include comments on items in your news feed,
What are your thoughts on the Facebook Lexicon? Do you think it may be useful in the future? Share your thoughts in the comments.
Currently the tool is in a test version that only has about 20 different topics to see data for, though once it is officially rolled out you will be able to search for data with any keyword.
Fortunately for those of us in the political realm, many of the provided keywords are political: McCain, Obama, Biden, Palin, Democrats, Republicans, and politics.
Here are some of the main features of the new Lexicon.
Dashboard
The dashboard allows you to look at the number and percentages of posters and posts. Above is the dashboard for Obama. You can see that on the bottom it breaks down the posters into age and gender demographics. Unfortunately this data is raw and not weighted, so the data seems to always resemble Facebook's demographics as a whole.
Demographics
The demographics tab will allow you to see graphs of the age and gender data. Once again the fact that you can't weigh the data diminishes a lot of its potential utility.
Associations
The associations tab shows you what other terms people use in the same wall post as the keyword. This is represented graphically using circles on an XY plot. The size of the circle represents the number of posts that included the associated keyword. The X-axis indicates whether more males or females use the two keywords together more frequently, with male being to the right and female to the left. The Y-axis shows the average age of the user making the association. I think this feature may become one of the more useful of the set, but the unweighted age data affects this as well.
Sentiment
The sentiment tab shows how often the keyword is in positive or negative contexts. The algorithm/dictionary they are using for this isn't public, so I'm not sure on the specifics of how they determine the score. It is important to note that there are a lot of false positives and negatives using these algorithms. They can't determine sarcasm, so if I posted "Yeah, McCain is exactly what we need to fix the economy" sarcastically on a friend's wall, it would probably show up as positive sentiment. That being said, you are still able to get a good general idea of sentiment by looking at the trend lines. In the screen shot above you can see the compared sentiment scores for Obama and McCain over the last 14 days.
Pulse
The pulse tab shows keywords that frequently occur in the profiles of users who mentioned the topic. You can see raw numbers for interests, music, movies, TV, and books. This is definitely more useful for commercial marketers than political people, and once again raw data means that something like Harry Potter is going to be at the top of all the book lists because of its ubiquity.
Map
The map tab shows where people are talking about the selected topic(s) and allows you to compare relative popularity. Each state is colored based on how many times the topic was mentioned relative to the other topic. In the screen shot above I compared Obama and McCain. In this case Obama is mentioned more than McCain in every state. If a state mentioned McCain more than Obama it would be a shade of red.
Final Thoughts
The Facebook Lexicon might be useful for some national campaigns in the future once it opens up to all keywords, but there are some things holding it back.
As I mentioned a bunch of times before, the fact that you can't weigh data is a big weakness. What I would like to see would be to compare the data relationship to the averages of Facebook users in the United States. For example, if 50% of Facebook users have Harry Potter as a favorite book on their profile, but 58% of posters of a keyword have it listed, show the data as +8%. Conversely if only 46% listed it, show -4%. This would allow you to get a better picture of the data relationship than the raw data does.
Another thing to keep in mind is that the Lexicon only uses data from wall posts. This does not take into account status updates. I am not yet sure whether or not the final version will include comments on items in your news feed,
What are your thoughts on the Facebook Lexicon? Do you think it may be useful in the future? Share your thoughts in the comments.
Add a Comment
Sign in
More Information
Pirillo's Posts
Lockergnome
© 2009 Created by Chris Pirillo
You need to be a member of Geeks! to add comments!
Join this social network