https://github.com/sunsided/coding-stats

Explorative data analysis of my coding stats
https://github.com/sunsided/coding-stats

c coding-stats cpp csharp eda python rust

Last synced: about 1 month ago
JSON representation

Explorative data analysis of my coding stats

• Host: GitHub
• URL: https://github.com/sunsided/coding-stats
• Owner: sunsided
• Created: 2023-07-28T16:48:59.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2023-12-02T16:11:26.000Z (about 1 year ago)
• Last Synced: 2024-11-09T08:25:17.387Z (2 months ago)
• Topics: c, coding-stats, cpp, csharp, eda, python, rust
• Language: Jupyter Notebook
• Homepage:
• Size: 1.79 MB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Exploratory Analysis of my Coding Stats

A simple exploratory analysis of my coding stats. I use WakaTime since late 2016 to track my

use of programming languages and periodically fetch the daily stats from the WakaTime API (see e.g.

my [coding-stats.json Gist](https://gist.github.com/sunsided/667936bd0d4d5f042ea4e6726c5faccb)). Since

recording only started in 2016, large chunks of C#, C/C++, Matlab and other languages are missing.

In this repo, I ran some trivial analytics from those JSON files. If for whatever reason you are

interested, the `redacted.csv` contains the language-specific information. I excluded other source

data from this repo since it can hypothetically contain mildly sensitive details such as project

and machine names.

I mainly wanted to see how my Rust journey went compared to other languages and at the very least

it's interesting to say that within the last two-ish years as of writing this, I spent as much time

writing Rust code as I spent time writing Python code since the recording started.

![](images/language_development.png)

The C# curve is pretty linear, which is not surprising since I have been working in .NET focused

companies since 2008 and therefore use C# a lot during my work life. What is interesting, however

is that for the last three years I have not actively assumed the role of a software developer

and _still_ maintained a linear progression there.

This also tells a story about COVID. During the period of 2020..2022, a lot of work was done from

home. The below plot shows the deviation of my actual C# coding times from a linear regression

model (R²≈0.99) in terms of cumulative sum of programming hours. As is obvious, this number rises

both in late 2019 and late 2020, to slowly and then drastically decline in 2022. If one were to look

for symptoms of a burnout, they're right there.

![](images/post_covid_burnout.png)

There is also an interesting apparent seasonality in the deviations that would be interesting to

analyze further.