Data Analysis project using twitter API and python

#Questions answered after the analysis:

1. Engagement with Dogs posts over time
2. Dogs rating over time
3. Which Dog breeds should WeRateDogs concentrate one ?
4. How active is WeRateDogs account ?
5. Which dog stage got the highest engagement
6. How reliable is the dog detection algorithm ?

The image shows the twitter account activity over the years

How active is WeRateDogs account ?

The Account owner is either losing interest or don't have the time to keep the momentum so tweets are down from 300 tweets/month  to only 50 tweets per month.

How reliable is the dog detection algorithm ?
Nearly three quarters of the 'dogs' in our master archive dataset are predicted to be dogs with some false negative in non-dogs, which is quite low for a dedicated dogs account. This puts prediction accuracy in question.

The Dataset was retirved from twitter using API , data was cleaned & wrangled to reach a few conculsion.

#The output of the wrangling process:

Two datasets tweet_archive_master with 1968 rows and image_predictions_df_clean dataset with 1971 rows.

They are saved to twitter_archive_master.csv and image_clean.csv respectively inside the data folder.