Inspiration

We noticed that some of the items listed in the data file resulted in a loss of revenue and were curious if there was a way to determine a more effective AWP to minimize losses.

What it does

The Python program parses the data file and then fetches the values that we felt had a correlation with AWP. The values are then fed to a regression model. A small portion of these values were used for training and the rest were used for testing predictions. It then continuously takes in hypothetical acquisition costs and outputs a predicted AWP value.

How we built it

The calculations were done in Python and a visual representation was shown in Excel.

Challenges we ran into

We felt that some of the data was inaccurately pre-calculated and that it affected the performance of the predictor.

Accomplishments that we're proud of

We were able to implement a machine learning algorithm for the first time.

What we learned

We learned about training models.

What's next for Pharmaceutical AWP Predictor

We may try to find more data that could improve the prediction model.

Built With

Share this project:

Updates