TweetTone shows positive online sentiments for African ICT, tech, internet, business
Ever wonder why that Tweet you deemed innocuous led to a backlash? Lymbix, a company working to better understand “sentiment analysis” has products that analyze written communication for emotional intent. The company currently offers two products: ToneCheck (used to prevent the sending of negative email) and TweetTone (used to analyze the emotion behind Twitter conversations). TweetTone is intended to be used in real-time by marketers to monitor the public sentiment behind their brands. However, a free service can also be used to gauge the tone of conversations around a certain keyword.
Based on recent Tweets from June 2-4, many African themes generate positive online conversation. A few examples are included below, but feel free to explore the tool for yourself.
“Africa ICT” evokes friendly and amusing conversation:
“Africa internet” generates enjoyable conversation, although there is a hint of uneasiness:
“Africa tech” is mostly positive (especially amusing), but 1/3 of conversations are deemed negative:
“Africa business” is over 75% positive:
“ict4d” takes the cake when it comes to contentment:
Of course, few algorithms are perfect. TweetTone appears fairly accurate over a large sample. Individually, however, emotion can be confused. Case in point is the recent @oafrica Tweet, “22 African ICT, mobile, broadband, and computer stories from the past 9 days http://bit.ly/li6fa0 #ict4d #africantech“. TweetTone classifies this middle-of-the-road Tweet as sad. I have no idea as to why. A similar Tweet mentioning genocide has been classified as sad, simply due to the inclusion of that word. The Tweet itself lacks much of a stance. Keep in mind:
- There’s no “right way” to measure sentiment. Some statements can have both positive and negative emotion, but end up ambiguous.
- TweetTone is only successful with the English language.
- The goal of TweetTone is to both be accurate with what a sentence should mean and agreeable as to how human readers interpret.