Because they can reveal the potential needs of users (customer), studies of silent customer analysis or silent majority analysis have increased recently. On the other hand, the user became able to post easily their preferences to the internet following the growth of the internet or SNS. This study proposes to uncover consumers' interests by user situation analysis using “boring” tweets.
We collected about 800,000 “boring” tweets between September 2015 and August 2016 using an independently developed script. Then, we calculated T values and MI values of these tweets, and extracted the parts of speech that co-occurred with the word “boring” (“hima”). Then, we revealed characteristics of the area where the “boring " tweets are concentrated by visualizing the location information of nouns which co-occurred with “boring” on a Japanese map. Furthermore, we revealed the user situation for example in what case does the user say “boring"? by analyzing deeply the nouns that co-occurred with the word “boring”.
View full abstract