Authors:
- Jan Trienes
- Anna Berger
- Andrés Torres Cano
Year: 2018
Abstract:
The domain of programming is dominated by partially self-taught developers. High industry demand for software professionals and excellent online-learning facilities cause many people to start a career in software development. However, as a newcomer to programming, it can be difficult to prioritize which technologies to learn as the space of possible options is vast. This research aims to visualize the dependencies and relations within the ecosystem of programming. Based on a dataset of 15 million questions asked on Stack Overflow, we create a weighted graph in which programming concepts have a notion of relatedness and belong to communities of programming concepts that share common properties. We furthermore demonstrate that the proposed method is also applicable to other problem domains based on a dataset of questions relating to statistics and machine learning.
Keywords: Graph Mining, Community Detection, Stack Overflow
Figure 1: Most popular tags on Stack Overflow (left) and most frequent tags co-occurring with iOS (right).
Figure 2: Graph of 256 most popular tags on Stack Overflow.
Figure 3: Graph of 256 most popular tags on Cross Validated.
Download