The purpose of this analysis was to learn how to use Google Colab, Pyspark, PgAdmin/ MySQL, and AWS to create data frames. The dataframes created in this analysis included data in regards to prodcut ids, ratings, reviews, and whether or not reviewers were in the vine program or not.
- There were a total of 64,968 non-Vine reviews and a total of 613 Vine reviews
- There were a total of 30,543 non-Vine five star reviews and a total of 222 Vine five star reviews
- There was a five star review percentage of 47% for non-Vine reviews and a five star review percenatge of 36% for Vine five star reviews
In conclusion, the five star review percenatge was lower for Vine reviewers as compared to non-Vine reviewers, so there is no proof of a bias for Vine reviewers compared to non-Vine reviewers. (As shown in the images above, the Vine reviewers 5 star percenatge was 36% compared to 47% for non-Vine reviewers) Another additional analysis that could be done to see if there is a is for Vine reviewers vs non-Vine reviewers would be to see the average star rating for both groups, to see if Vine reviewers are more prone to give a higher rating as compared to non-Vine reviewers. This could also be taken a step further to see the percentage breakdown for each star rating possible (i.e. 5 star vs 4 star vs 3 star etc.)