Giters

ugis22 / analysing_twitter

Stream Tweets and store them in a relational DB. Perform sentiment analysis and network interaction.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Question: tweets_predict_* vs tweets_*

Nowi5 opened this issue · comments

commented

Dear ugis22,

thank you very much for your work. I try to follow but I have a question in mind.
So far I understood, you listen to Twitter API and every new Tweet as from interaction or from keyword goes into table predict. In Database clean etc. you use both tables. So I would assume that you have filled tweets_* beforehand. This script is missing in here, right?
Why did you separate the two tables?

Thanks

commented

What I did is stream one time for the training part, and store it under the table tweet_avengers. Then, stream a second time and store it under the table tweet_predict_avengers. The cleaning DB is because I do not want any duplicate tweet to be in my tweet_predict_avengers, as I need only tweets that haven't been seen by the model before.