mattarderne/firefox_explore

Explore your firefox browsing history trends using Metabase Analytics Tool

firefox_explore

be-true-to-thine-fox

Explore your firefox browsing history trends using Metabase, an open source analytics tool

Disclaimer - this was made in a rush, and sat on ice for a while, please PR any suggestions & glaring oversights.

Background reading on the Firefox ERD can be found in the rather sparse Firefox docs (the ERD is no longer valid)

Setup

Clone this repo

git clone https://github.com/mattarderne/firefox_explore.git
cd firefox_explore

Find where your Firefox places.sqlite file lives in about:support page under the Profile Folder entry and copy here.

cp ~/Library/ApplicationSupport/Firefox/Profiles/y4pw28fm.default/places.sqlite .

The below creates a docker container, linking this repo (with the Metabase backend sqlite database and the Firefox places.sqlite) to the Docker container

docker run -d -p 3000:3000 \
-v $PWD:/metabase-data \
-e "MB_DB_FILE=/metabase-data/metabase.db" \
--name metabase_ff metabase/metabase

Run below and 2-3min later Metabase will be running at http://localhost:3000

docker start metabase_ff

Use

Login details are:

admin@admin.admin
admin11

Links

• Browsing Overview, use as a starting point
• Base SQL table, use this as a base for new queries

The dashboards sometimes need to be refreshed after their first run if any questions don't load

Procrastinate

This allows you to define a "bad list" of sites that you'd like to mark specifically (or use as a way to insert custom data into the sqlite db)

1. Modify the site_list in procrastinate.py with the domains you want to include, use this question to get your top 20 list.
2. Run python procrastinate.py to copy the lists to the places.sqlite

TODO

Contributions useful, the following are necessary at some stage, but other useful things include creating useful visualisations (kinda tricky to integrate as the code lives in the metabase.db), cleaning up the all_site_visits.sql file

Small

• figure out your most clicked HN titles, look at the words in the title
• Name it something clever - true-to-thine-fox
• Fix the Docker to use docker-compose
• Mount the places.sqlite directly in place rather than copying it to the repo
• Find a better way of doing procrastinate_base
• think about doing an integer join on the all_site_visits table on the LEFT JOIN procrastinate on cleanup.top_level_domain = procrastinate.procrastinate join

Big

• Fix the dates in the base Base SQL table
• python, add some deep learning correlation graphs, jupyter notebooks, streamlit
  • ai on docker
  • streamlit docker
• find some other sqlite data sources and add them