Predict-Spotify-Top200

A report on what we have done:

results/report.pdf

Clone it:

git clone https://github.com/hd4niel/Predict-Spotify-Top200.git

TODO

• Analyze features' correlation with streams and position
• Timeline of features over time
• Feature timeseries forecasting
• Extra: Generate lyrics

File structure

data

Data files, .csv etc...

notebooks

Jupyter notebooks

src

Code, .py files

results

Analysis docs

Example workflow using virtual environment in python

Initialize virtual env in venv/ (it is ignored by gitignore)

virtualenv -p python3 venv

Activate virtual env

source venv/bin/activate

Install dependencies

pip install -r requirements.txt

Install package

pip install package

Add to the requirements

pip freeze > requirements.txt

Deactivate environment

deactivate

Use jupyter

ipython kernel install --user --name=.venv
jupyter notebook

choose kernel .venv

Notes

Used ImageMagick in bash for GIF

convert -delay 15 *.jpg topfeatures.gif

Feature explanations

Explanation at spotify dev site