NBA Draft Success from 2005-2015

Project Report

---

Extract

The data we utilized was from the following sources: Data.world, basketball-reference.com & Kaggle. The data we had access to was in csv format for from Kaggle and Data.world. On basketball-reference.com there was the option of gathering data via csv but we needed data from several years and the csv’s didn’t provide the team or year. Because of this we wrote a for loop that built a custom url for each year (2005-2015) for each finals team and grabbed the table, added team and year and created a dataframe for the winners and losers.

Transform

Then both dataframes were merged into one. This dataframe was then exported to a csv and a database was built in postgresql with the data as well as the draft data we gathered. Once in the the database we created a jupyter notebook to transform the data.

Load

With all the data and visualizations created we built a webpage using bootstrap. In our Postgresql database we had 2 tables (draft & finals_roster). With that database we were able to create the distribution graphic of all the draft rankings for all the finals rosters and see how they distributed. We also used it by joining both tables to see what the average draft position each position had for all the players in the finals during the period queried.

We chose the period of 2005-2015 because we felt a 10 year period would be a good range for this analysis. With more time we could look at other 10 year periods and see if we see any interesting correlations.