Analysis of Python Developers Survey 2019 Results collected by JetBrains.
Raw data can be found here: https://drive.google.com/drive/folders/1GW86M_QLFOA2KyeWeb8hzBj_RqQIXmAU?usp=sharing
Project was performed as a demo and this restricted the work to a < 1 week timeline. The demo was to show how you might perform and present a cluster analysis. This was the sole reason that a cluster analysis was performed over other techniques.
Questions were analyzed in 3 groups: (1) how involved are you with, (2) general, & (3) how often do you.
If these clusters were to be used to segment PyCharm customers. The main groups might be:
- Less experienced programmers, interested in more Data Science focused features (Cluster 3)
- More experienced programmers, interested in more traditional Software Engineering features (Clusters 1 & 4)
- Python web developers that are likely more interested in tools for building web applications (Cluster 0)
Analyzing / polling these groups might lead to valuable market analysis of these segments.
- Our most experienced cluster is the most involved with varying techniques - Our least experienced cluster is the most involved with more "Data Science" techniques - Cluster 0 doesn't stand out in many ways other than their preference for web devOrder of average coding experience: 4, 1, 0, 2, 3
Order of average Python experience: 4, 0, 3, 2, 1
- Cluster 1 had the biggest change in experience compared to general programming experience. (they dropped from 2nd to last)
- Cluster 3 had the 2nd biggest change in experience compared to general programming experience. (they jumped from last to 2nd)
Might be a group of Python web developers. Could be of interest to see this cluster's affinity for flask / django / etc.
Python is not their first choice for programming. They are experienced programmers, and they use varying technologies and techniques
Less experienced programmers with a focus on data science style tasks. Is their lack of programming experience due to a lack of experience in industry or due to a heavier focus on the math/stats of data science tasks?
Most experienced Programmers and most experienced Pythonistas. They are more classical programmers than they are data scientists.
Might not be a cluster of interest. Membership seems mostly based on not answering the "How often do you... ?" series of questions.
- This analysis ignored many of the survey participants and many of the questions.
- These decisions were made due to a short project timeline.
- Treating a lack of response as an indication of not using a technology / technique should be revisited.
- There is survey data available for 2017, 2018, & 2019. Only 2019 was considered. Could previous results be leveraged in analysis?
If these clusters were to be used to segment PyCharm customers. The main groups might be:
- Less experienced programmers, interested in more Data Science focused features (Cluster 3)
- More experienced programmers, interested in more traditional Software Engineering features (Clusters 1 & 4)
- Python web developers that are likely more interested in tools for building web applications (Cluster 0)
Analyzing / polling these groups might lead to valuable market analysis of these segments.