Scripts to parse and analyze Google Voice data exports.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.

Google Voice Analysis

A collection of scripts to parse and analyze exported Google Voice data.

So far it just parses the data and runs a few Naive Bayes classifiers, but I'm looking to expand.


  1. Download and extract your Google Voice data.

  2. Install the dependencies with something like pip install -r requirements.txt.

  3. Copy to and configure appropriately.

  4. Run to load everything into a SQLite database.

  5. Run to run a few simple analyzers using Naive Bayes.


  • people_with_many_texts(n) classifies the texts of people who have sent more than n texts.

  • recipient_is(name) classifies texts into either name or not_name.

  • split_me_not_me() classifies texts into either me or not_me depending on the sender.

Interactive Mode

  1. Get your training and test sets from one of the classifiers.

  2. Run build_classifier(training_set).

  3. Run interactive on your classifier.

  4. Enter some messages.


MIT/X11 licensed, except for and and are copyright Brendan O'Connor, Michel Krieger, and David Ahn and licensed under the Apache License 2.0.