stackoverflow_survey

A data analysis project using 2017–2019 StackOverflow Developer Survey Results.

Project Motivation

The purpose of this project is to implement the CRISP_DM process and data wrangling skills with real-life datasets.

Dependencies

• python 3.7
• jupyter notebook
• numpy
• pandas
• matplotlib

Project Structures

- dataset/
  - developer_survey_2017/
    - survey_results_public.csv
    - ...
  - developer_survey_2018/
    - survey_results_public.csv
    - ...
  - developer_survey_2019/
    - survey_results_public.csv
    - ...
- StackOverflow_Survey.ipynb

dataset/: A folder used to store 2017–2019 StackOverflow Developer Survey Results. NOTICE: Due to GitHub's 100 MB push limit, you will have to download the data from here, extract the archives and place them following the project structure above.

StackOverflow_Survey.ipynb: A jupyter notebook inluding all necessary project code.

A Brief Summery of Analysis Results

1. We counted the proportion of individuals working with particular languages , and compare the rankings in each year. It showed that the most popular language has been JavaScript. Up-trending languages include Go, Rust, Clojure, F# and Erlang.

2. We then looked at how different languages relate to job satisfaction. We found that TypeScript and Bash/Shell have good rankings in terms of job satisfaction these years. Steady increase in job satisfaction was observed in Clojure, Ruby, Go, R, Rust and Erlang.

3. Finally, we investigate which languages most people choose as their starting points by counting the proportion of individuals with a 0–5 years of coding experience. The top-5 languages in recent years are Assembly and Matlab(2017–2018).