Real time #ebola chatter scanner

I started reading every #ebola tagged wordpress blog post in August of 2014. But By October there were over 200 posts a day, so I wrote a simple chatter scanner. Every ten minutes it pulls all #ebola posts and saves them to a database that feeds this website:

When you load the page, it visualizes the last 24 hours of blog activity:

ebola-wordtree-oct-18Each word in the wordtree represents a theme in multiple blog posts. What surprises me a little is that the screen shot from the last few hours has the phrase ‘gun control’ in it. It shows that certain fearmongering blogs conflate the threat of ebola with other threats they harp on, like the right wing fear that Obama is going to ‘take our guns away.’ Algorithms do not exaggerate. When someone tags a post with #ebola and uses the phrase “gun control” repeatedly in it, the map shows.

What I wished I could do was write an algorithm to extract only first hand accounts from people directly affected by Ebola in West Africa. In my manual counts, some days there isn’t a single authentic narrative in over 200 posts, so I have no hope to writing an algorithm to find something that doesn’t exist. But I will keep looking for the voices of the real victims among the sea of suspicion and blabbermouthing on the Internet.

The minimum viable version is up. Better tools will follow.