README.md

deezer-stats

Visualize your Deezer listening habbits with this project.

Feature

Export to an Elasticsearch stack to use Kibana's visualization mecanism. Use the power of Python to get stats about the most listened artist, song, etc.

How to use

Elastic stack

Deploy the local Elasticsearch stack with a simple docker-compose up -d Aims your browser to localhost:5601 to access Kibana's UI.

Prepare your data

Get your Deezer data by making a GDPR request to their privacy email, asking specifically for the whole listening history.

They'll send you a big .xls containing everything. Export the listening history tab to a TSV file (using google's sprreadsheet, excel or whatever).

The support response may take some time, be careful (legally at most 31 days)

Run the script

Prepare your python venv, and simply run ./main.py <tsv file>

Import pre-built dashboards and vizu

gunzip export.ndjson.gz and import the ndjson file in Kibana's UI

Example

CLI

Get help

./main.py -h

Get your all-time top 50

./main.py --top-track --top 50 ./history.tsv

Get all the songs you've heard at least 100 times

./main.py --min 100 ./history.tsv

Get all the songs you've heard at least 10 times in decembre 2017

./main.py --min 100 --date-after 2017-12-01 --date-before 2018-01-01 ./history.tsv

Determine that a song heard is a song listened at least 30s

./main.py --top-track --valid-algo min_30 ./history.tsv

Get the song you've spent the most time listening to

./main.py --top-track-by-time ./history.tsv

Get the most listened artists

./main.py --top-artists ./history.tsv

Get the most listened artists by time

./main.py --top-artists-by-time ./history.tsv

Get the most listened artists by time and considered different artists for comma

./main.py --top-artists-by-time --split-comma-separated-artists ./history.tsv

For example, the results will be:

[  20]  929 times (2 days, 8:18:22 ): Claptone - from 2016-05-19 23:06:06 to 2020-10-18 23:14:11

instead of

[  41]  514 times (1 day, 4:48:43  ): Claptone                   - from 2016-05-19 23:06:06 to 2020-10-18 23:14:11
...
[  47]  355 times (1 day, 0:30:47  ): Claptone, Nathan Nicholson - from 2017-02-18 14:08:16 to 2020-10-18 22:41:20
...

Find the songs you liked back in 2017-01 but are not in your all-time favorites

./main.py --forgotten-hits \
  --forgotten-hits-start 2017-01-01 \ # Set the start period. Defaults to last year
  --forgotten-hits-end 2017-02-01 \ # Set the end period. Defaults to now
  --forgotten-hits-top 150 \ # Will yields the songs you liked the most that are not in your TOP 150 all-time
  --forgotten-hits-bucket-size 10 \ # Will give you 10 songs per month in the given period
  ./history.tsv

Kibana's UI

Heatmap of when I listen to music in average during the week.