Find query results here
The purpose of this analysis is to extract , transorm and load data based on the reviews collected by Amazon for their Wireless devices by using PySpark to load the data into pgAdmin via AWS RDS.
- Based on calculations the number of Vine and non-Vine reviews are as follows:
-
Number of Vine reviews
-
Number of non-Vine reviews
- The number of 5 star Vine and non-Vine reviews are as follows:
-
Number of 5 star Vine reviews
-
Number of 5 star non-Vine reviews
The percentage of 5 star Vine and non-Vine reviews are as follows:
-
Percentage of 5 star Vine reviews
-
Percentage of 5 stay non-Vine reviews
The results of this analysis clearly indicate that there is not a bias towards Vine reviews or 5 star Vine reviews as it is clearly visible that there were more non-Vine reviews tabulated for boh categories . This is also true for the resultant percentages where a larger percentage of 5 star reviews came from non Vine program participants(43.62%) vs Vine reviews at (36.08%). A futher analysis that could be useful is to use the NaiveBayes model to attempt to predict the likelihood of review being written by a Vine vs non Vine user .
See prediction here