Yahoo Japan have released a new sentiment analysis function in its real-time search results in this month.
The development, although in its beta stages, mechanically scans tweets in real- time search for positive words such as, “goodbye”, “thank-you” and “awesome”, etc. It also scans terms with negative connotations such as “bad”, “upset” and “angry”.
After searching for a keyword in Yahoo!Japan’s real-time search, a pie chart will be displayed showing the proportion of tweets mentioning emotive or sentimental content and will classify them as positive or negative.
The sentiment analysis only scans for popular keyword terms, both positive and negative, and therefore currently does not display unpopular keywords in the sentiment results. The update is available for both desktop and mobile versions of Yahoo!Japan.
To simulate how the sentiment analysis works, we reviewed airline- related keywords. For ANA, it appears that more tweets using the English name “ANA” show positive sentiments compared to the Japanese name “全日空”:
Only domestic Japanese airlines (JAL and ANA) have sentiment analysis displayed in the results, Yahoo only produce the analysis for keywords associated with higher amount of tweets and mentions. This may however change after the beta phase so less popular keywords will also be scanned for and presented in the results in the future.