caseyzhang11 / ec601 Goto Github PK

Your project is really great, impressive.

It is well-designed and precisely implemented. You designed two components, and in the two components, you designed classes and methods. The structure is clear with modulars. And it works well, exactly achieves your MVP of extracting and analyzing Twitter trend sentiments.

Your code is clear with great readability, because your variable names are self-descriptive, and the code is well-commented.

You also have great error handling in your TWITTER STREAM LISTENER. It handles errors that may occur in printing data and writing data to fetched_tweets_filename.

You also have clear documentation with examples of extracting tweets and analyzing sentiments.

You also did unit test on possible errors users may meet with, which is great. You tested what will happen when Twitter credentials are wrong and also, when Google Cloud NLP service is down.

Your project will be even better if:

1. Not putting your Twitter credentials on GitHub, because it may be abused if someone stole your credentials.
2. Integrate your two components, to make it more easy to use for common users. Maybe user only need to input keyword and will get results (tweets extracted and their sentiment) automatically.
3. I'm not sure if it's my problem, when I run TwitterAPI.ipynb on Google Colab, it prints huge amount of data, and it seems that it is running forever, making the page really slow. Maybe you can offer users an option to extract tweets of the latest several minutes, or limit the extracting speed, rather than lively extracting all tweets about the keyword