Scripts and functions to scrape and analyze text from ASAP Sports, a sports interview transcript site. Thus far the code is focused entirely on practice day interviews between National Hockey League Stanley Cup Final games. The initial study of the data can be found in this Analytics Vidhya Medium post. I've published the dataset and an introductory notebook on Kaggle. In the introductory notebook, starter_notebook.ipynb, I contrast interviews with coaches and players and build a simple TFIDF-based model that determines whether a given interview transcript is coming from a coach or a player.

The functions (in bold below) that collect, clean, and access the data are well commented and written with other users in mind.

starter_notebook.ipynb:

• explores and introduces the data, creates and tests a coach vs player classifier

create_csv.py:

• converts the html-tagged data into a Pandas dataframe, which is then saved as a csv

scraper.py:

• scrapes interview data from the hockey portion of the site

clean.py:

• corrects inconsistencies in the website's data

utils.py:

• helper functions for gathering the scraped data

plotting.py:

• creates figures that study these interviews

data/interviews_raw.txt:

• the scraped data

data/interviews_clean.txt:

• the scraped data, cleaned using clean.py

data/interview_data.csv:

• the data from interviews_clean.txt, in a csv format

Several of the figures contained in this repo are shown below. They are explained in detail in the Medium post .