Master's Thesis - Quantitative quality assessment of open source projects

Research and develop a model to estimate the quality of an open source project!
Scroll to content

that takes none-community data into account 

Background

Evaluation of open source projects are currently performed manually in the industry, which makes the results susceptible to subjective opinions and error due to variations in experience of the evaluator. Debricked has developed a model for quantitative evaluation of open source projects that creates metrics in terms of contributors, community, security, and popularity, but would like to expand these metrics with a measurement of “quality” as well. For this project, we would like the student(s) to research what quality aspects are important to open source consumers, such as test-coverage, quality of documentation, quality of code, complexity, ease of use etc. The student(s) would go on to implement the model and derive the quality metric for the hundreds of thousands open source projects monitored by Debricked. To analyze the quality of the metric, the students would need to use sophisticate statistical modeling to estimate the accuracy of their metric in an unsupervised manner.

Project

In this project, the students would research and implement an open source quality metric. The metric would be developed to as accurately as possible represent the underlying quality of the software, while still maintaining high transparency and the ability to generalize to as many types of open source projects as possible. The steps of the project could be:

● Determine what quality aspects of open source projects are important to open source consumers through a literature study and interviews.

● Implementing a data-pipeline to efficiently extract relevant open source data, such as code, inline comments, and documentation.

● Developing a model, the estimate the quality of an open source project.

● Performing statistical analysis to determine the accuracy of the model with unsupervised methods and latent variable error estimation.

Python will be the chosen programming language for the implementation of the proof-of-concept in the project.

 

The project is suitable for 1-2 students with interest and background in software engineering and programming. Knowledge of open source, python, databases, software practices, and statistical analysis is very relevant to this project. 

Usually responds within a week

Or, know someone who would be a perfect fit? Let them know!

Minc, Malmö

Anckargripsgatan
211 19 Malmö Directions View page

Why join Debricked?

  • Startup

    Working at a startup is a great place to build upon your existing skill sets, gain experiences in many different areas, and take a ton of responsibility. As we grow quickly, so will your opportunities for career advancement.

  • Cutting Edge

    We are building cutting edge algorithms, AI and other buzzy things with patents filed. The challenges and opportunities at a startup constantly change which means you are forced to adopt new skills, take on new projects, and will never be bound to a dull, boring routine.

  • Team Spirit

    Professional when facing customers, but in house we are a mostly freshly baked graduates with a playful and creative mindset.

    We move fast from creativity to decision to implementation and support each other to be as good as we can be. You will have the ability to impact the team with your thoughts and resourcefulness.

Workplace & Culture

Debricked offers a real startup culture that values your innovativity, creativity and ability to take initiative. It is all about taking action on your ideas and getting things done to shape the company together.

Teamtailor

Applicant tracking system by Teamtailor