Software Engineering Intelligence (SEI) refers to the use of advanced analytics, machine learning algorithms, and data science techniques in the field of Software Engineering (SE). The primary goal of SEI is to enhance decision-making processes, improve software development methodologies, and optimize project management practices. Traditionally, SE relies heavily on expert opinion for decision-making. Using SEI, development organizations can increasingly make informed decisions rooted in data.

CodeScene is a modern SEI platform that blends static code analysis with insights into organizational dynamics. Unlike traditional tools that focus solely on code quality, CodeScene also provides insights into the human aspects of software development. The tool leverages the rich data from git logs to reveal organizational and collaborative dynamics. Core features of CodeScene include technical debt management and software visualization. The visualizations illuminate aspects crucial for engineering management, including code maintainability, developers’ knowledge distribution, and off-boarding risks.

For developers, CodeScene provides assistance through automated code reviews and quality gates in the CI pipeline, i.e., providing code smell detection in pull requests. Moreover, CodeScene provides an IDE integration for VS Code that help developers mitigate code smells before committing code. Recent R&D efforts have increased the level of automation in IDE-based code smell management. Instead of only identifying code smells, CodeScene now proposes refactorings based on input from Large Language Models (LLM). As LLMdriven refactorings are plagued by uncertainty, a fundamental concept is to provide guardrails and reject invalid suggestions. This feature is currently in closed beta testing with selected customers.

We will demo the commercial tool CodeScene. The tool is based on both SE and psychology research. Hopefully, tool-oriented conference participants will find it interesting to see what a mature SEI platform can look like.

Example screenshots are available on the company webpage: https://codescene.com/