We know that sometimes it is complex to check  queries and even ,more difficult to see if they are correct or wrong, for that reason we recommend you to follow the next steps.

1.-  When orchestrating the query check the operators that are available . Remember that the query needs to have less than 500 characters!
2.- Verify the query in Twitter before scraping in Tractor. (you can see wether you are obtaining the expected results or not.
3.- After everything works well in Twitter, scrap data from Tractor. If you are experiencing some inconsistencies in the data, get back to us for help.
4.- Create a project in Graphext. If the project contains more than 50k tweets, we recommend to use create a project using new recipes as it will allow you to sample your dataset:

  • Random sample: it will randomly select tweets.
  • Top/bottom results: you can choose the top results based on a variable or column (example top retweeted ones)
  • Filter using numerical selection
  • Filter using categorial selection.

5.- Analyse your graphext and obtain nice Insights.

TIPS: whenever you create queries you will need to consider the noise that will create, therefore it is essential trying to reduce that noise to the maximum in the query. 

Did this answer your question?