Skip to content

Commit message analysis

Silvio Montanari edited this page Jan 1, 2017 · 4 revisions

Description

An often neglected piece of information available through the version control history are the commit messages associated with each revision of the codebase. These messages, to a certain extent, describe the reasons behind code changes and can provide useful insights on how the software development process relates to the business processes, both from a technical as well as a social point of view.

Tasks

commit message analysis

This task captures the number of occurrences of any word in the all commit messages within a given time frame.

Usage
$ gulp commit-message-analysis [--dateFrom=<dateFrom>] [--dateTo=<dateTo>] [--timeSplit=<timeSplit>] [--minWordCount=<minWordCount>]
  • The minWordCount parameter can be used to set the minimum number of occurrencies for a word to be included in the report.
  • The timeSplit parameter again it's used to split the report in multiple temporal periods and observe changes in time of the most frequent words.

The commitMessageFilters configuration settings also can be an effective method to minimise the noise created by words not very meaningful but with a high occurrency rate (e.g. determiners, conjunctions, prepositions, etc.). See the Configuration for details and examples.

Visualisation

Results of this report are displayed through a well known word cloud diagram:

Word Cloud diagram

A filter feature is available to only show words with the most occurencies:

Word Cloud diagram - filter

Clone this wiki locally