A data analysis project using 2017–2019 StackOverflow Developer Survey Results.
The purpose of this project is to implement the CRISP_DM process and data wrangling skills with real-life datasets.
- python 3.7
- jupyter notebook
- numpy
- pandas
- matplotlib
- 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.
-
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.
-
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.
-
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).