This is my capstone project on experimental design for Thinkful’s Data Science Flex program. For this capstone we were to complete a buisness research project. Steps involved
- Write a research proposal
- Run a hypothesis test, analyze and write up results
- Create a power point presentation
- The research proposal, https://github.com/mathisme/ThinkfulDSCapstone1/blob/main/Thinkful%20Research%20Proposal.pdf
- Jupyter notebook containing the results, https://github.com/mathisme/ThinkfulDSCapstone1/blob/main/Results.ipynb
- PowerPoint presentation, https://github.com/mathisme/ThinkfulDSCapstone1/blob/main/ThinkfulDSCap1.pptx
As someone who uses apps and gets annoyed by all the advertisements popping up in free apps, I was curious to see if free ad-enabled apps get lower ratings on average than paid apps without ads. To examine thiis, I chose the Google Play Store Apps dataset found on Kaggle. The dataset can be found here: https://www.kaggle.com/gauthamp10/google-playstore-apps Although this is not for any buisness purpose, developers can see from my research that having ads does not significantly effect user ratings.
- An exploratory data analysis was performed on the data
- As a t-test was to be used, the ratings column needed to be transformed to make it more normal
- A/A testing was used to test for bias
- A one sided t-test was performed and results analyzed
- To derive a confidence interval on the non-transformed 'ratings' column, bootstrapping was utilized
- Numpy
- Pandas
- Scipy
- Matplotlib
Find data on user reviews and see if ads effect review sentiment