- Use Twitter stream api and Tweepy crawl 50000 real-time twitter with the key words related to sneakers.
- Divide the Twitters into different groups based on key words. Use pie chart to show the ratio of different brands.
- Clean twitters and find the important words in each groups. For example, lots of twitters in Adidas group mention Utd Man because the team just won a champinship of Football League Cup. At the same, we can find Yeezy and boost are the most popular shoes now because they occur a large number of times and distribute in different groups.
- Use Word Cloud to visualize the text in different groups.
- Get Amazon review dataset with reviews and ratings. Use this dataset to build a sentimental analysis model with TF-IDF.
- Use the sentimental analysis model to analysis the tweeters in differet groups and find out the distribution of positive, neautral and negative twitters. Use stacked bar chart to visualize the result.