Replication Package This section briefly describes the contents of the replication package, which makes it possible to replicate this study. Dataset construction: The recreation of the GitLab AndroidTimeMachine instance is done using the snapshot related code found in the Kotlin program. This program uses a running instance of the AndroidTimeMachine’s Neo4j database and uses its data to make API calls to a running GitLab instance which imports the GitHub projects found. This step can be skipped, and a GitLab instance can be instantiated directly using our created Docker volumes. The is also capable of performing the Kotlin filtering steps. For this, it uses the results of the previous steps as input and clones each repository for performing the filtering procedure. SLOC counting: SLOC counting related scripts are found in and implement our methods by cloning the project and running CLOC. It also plots the line graphs that we manually categorized into evolution trends. The raw SLOC data is found in , and all of the plotted evolution trends are found in. Kotlin migration mining: The implementation for mining is located in , and implements our methods by cloning a repository from the running GitLab instance and executing described methods. All commits we found are located in. Run-time efficiency: As mentioned in the thesis, we used AndroidRunner. A fork of AndroidRunner we used for conducting our experiments is located in . In the folder , each configuration combined with raw results is found per test subject. The final data we used after removing corrupt frame times is located in the folder, which also contains . This python script implements our statistical analysis methods.