Using sentiment analysis on Tweets streamed for 7 days (from February 1 2020 to February 8 2020), I studied the reputation (polarity) of Netflix and Disney+ in the US and Europe, and used Bokeh library to visualize these results.
This repository includes :
- tweets.ipynb : contains the main code for streaming and processing tweets to get the polarity of each tweet.
- tweet_app.py : contains the code for Bokeh application.
- df.csv and df_prepared.csv : contains streamed data.
Note that I was only able to stream Tweets for 7 days. You could use another hashtag in tweets.ipynb instead of Netflix and Disney+ to compare other companies, or even other topics.
pandas, tweepy, bokeh, nltk and cryptography
To visualize the results of this projet, loacte to this repository after cloning it and execute this command :
bokeh serve --show tweet_app.py
Here is the visualization of polarity of Netflix and Disney+ tweets in both US and Europe.
- Europe :
- USA :
Despite the limited data (only 7 days), I was able to get some useful information from it:
- Disney+ is more popular in this period than netflix in Europe. In fact, during this period, Disney+ reached its peak in the stock market.
- On February 5, 2020, Toy story 4 was released on Disney+, which justifies the increase of positive tweets on Disney+ in the US compared to netflix, which started to decrease.