In this first phase of the study, we analyze tweets from Bangladesh for a period of 6 months. We extract keywords and explore their frequency. Then we move on to our subject matter and focus on tweets related to woman/women. We explore the keywords associated with the words woman/women. Finally, we produce a visualization of the keywords and their frequencies in the form of word clouds. The output of this analysis is as follows:
- Top 100 keywords from the tweets.
- Top 100 keywords associated with the word 'woman'.
- Top 100 keywords associated with the word 'women'.
- Word cloud visualization of all tweets.
- Word cloud visualization of tweets that contain the words 'woman'/'women'.
- A visual comparison of the proportion of tweets that are focused on the subject matter.
The duration of the collected data coincided with the US election period. That explains the high association with words like Trump and Hillary. Words like men, man, love, beautiful have come up quite frequently. It is important to note that words like attack, rape, violence, nasty, abortion, etc have also come up quite frequently in these tweets which hint to the tensions in gender issues in the region.
Use of twitter is fairly limited in Bangladesh. The portion of Bangladeshis that do use twitter are from a very narrow cross-section of the society, mostly the urban upper, upper-middle class. That explains the fairly modest amount of tweets regarding the subject matter. Focusing on facebook may yield better discovery in this regard as facebook has penetrated nearly all social classes in Bangladesh.