Predicting claps from headlines on Medium.
Due to the wide variety in options for managing virtual environments, we leave it up to the user to create and activate your virtual environment.
First run to setup Python packages, .env
& folders:
$ make requirements; make dotenv; make clean
Download data & model artifacts into ~/hacking-medium-headlines
from S3:
make pulls3
$ python3 src/cli.py
Download data into ~/hacking-medium-headlines/data/raw
:
$ python3 src/scrape.py
Run grid searches and train final models:
$ python3 src/ml.py
Also included are notebooks documenting part of the process
notebooks/eda.ipynb
- initial data analysisnotebooks/error-analysis.ipynb
- looking for patterns in the errors made by the model
$ pytest src/tests.py