ssw-565-final

Author

Stephanie McDonough

Description

Objective: Automate and optimize the code review process in OpenStack repositories by categorizing check-ins based on their architectural attributes.

P1. Code Check-in Clustering:

• Export a list of more than 500 code check-ins from OpenDev Links to an external site.
• Cluster the check-ins based on architectural attributes (e.g., documentation, performance, security).
• Analyze the statistics of the clusters, their percentage distribution, and the challenges in detecting certain clusters.

Method

Used KMeans clustering to categorize code check-ins from OpenDev into architectural attributes like performance and usability. Pandas and numpy were used to statistically analyze the clustering output and matplotlib was used to visualize percentage distribution.

Results

1. See CodeClusters.py and figs folder for general code check-ins data visualization and analysis
2. See SecurityClusters.py and security-figs folder for security-specific code check-ins data visualization