Some people when they get bored, they go to the mall or play video games. For me, my solution to boredom is to write code unrelated to my current job. I admit this is not exactly what most people’s definition of fun is.
Anyway, last night while I was scanning Twitter for something interesting, I noticed a lot of posts about the Philippine presidential election, which is less than 2 months away. I guess things are heating up en route to the big game. Hmmn. I been studying data visualization recently, so why not take a shot at these Twitter posts and see what I can find. After a few hours of trial and error, I hereby present a visualization of the Philippine election based on Twitter posts.
Step 1: Click the button below:
Step 2: The fun part
Type keywords ‘Villar’ (without the quotes) and press Enter
You can enter 1 or several keywords and it will show you the context in which those keywords appear. For example, if you enter ‘Ruffa’, you will see the image below. I don’t know the issue between Ruffa and Noynoy Aquino but apparently her mother (as usual) is involve.
When I tried ‘Kris’, I learned that Noynoy’s survey ratings dropped. OK, enough with the spoilers! I suggest trying out different keywords and share what you have discovered.
You can also click on the keywords on the chart. This is cool for exploring the different posts available.
Now, for some geekspeak.
The chart you are using is called a word tree. A word tree allows you to search for words or phrases and present it to you it a tree-like structure. Notice some words are larger than the rest. The size of the word represents how many times the word was found. In essence, the largest word is the most talked about topic related to Villar and Noynoy during the past 4 days.
The cool thing about Twitter is it is programmatically easy to get tweets by people a few days back. First, I tried to pull tweets as far as Jan 1 but unfortunately, Twitter has an artificial limit of 1,500 tweets only. As of this writing, I was able to collect posts between Mar 7-11 only. The date range wasn’t intentional by the way.
The charts was created using the Many Eyes website – a visualization project sponsored by IBM Visual Communication Lab. You can try their visualization tools for free and all data sets are widely available. You can find my Twitter data sets here.
I will continue collecting data until election is over (or maybe until the new president is proclaimed). I am thinking of creating charts on a weekly basis instead of lumping everything because I think it is fruitful to get a visualized snapshots of the election on a weekly basis.
What do you think? Please leave a comment for any suggestions and violent feedback 🙂
Update: The chart doesn’t load on Chrome/OSX but works fine in Safari and Firefox (both on Mac). Sorry, I haven’t tested it in IE and no plan to do so 🙂