citizen science – open data – Python

Can I Still Code?

How is vibe coding changing our Python skills?

A longitudinal project tracking real developers over time

Join the Study Learn More

The Mission

We're entering an era of agentic coding, where tools such as Cursor, Claude and GitHub Copilot become commonplace.

But what impact does this have on our coding skills? Does it free us to solve higher-level problems?

Can I Still Code is a longitudinal, citizen-science research project designed to measure exactly this.

Ethics approved by Royal Holloway, University of London Research Ethics Committee — Ref: 991

Discovering what protects your skills

Beyond measuring coding ability impact, this project also explores what protects your skills, despite heavy AI use. The plan is to share identified protective habits back with the whole community.

How It Works

1. Register

Create a profile and provide your baseline demographics and coding habits.

2. Assessment

Complete short Python challenges periodically (no AI, no Googling).

3. Insights

Explore your personal performance trends and see how you compare.

4. Longitudinal

Return every 28 days (but just 2 sessions a year apart is useful!) to build a robust dataset of your skills over time.

Try a challenge

Click to try a challenge

Loads in ~5 s on first run

Challenge: sum_of_digits

Write a function sum_of_digits(n) that takes a non-negative integer and returns the sum of its digits.
Example: sum_of_digits(123) → 6

Python 3.x

How they work →

Your insights panel

Your trends you have access to.

Accuracy (demo) ⓘ

Accuracy percentage over time.

Time Taken (demo) ⓘ

Average time per challenge in seconds over time.

Submissions Per Challenge (demo) ⓘ

Average runs per challenge over time.

Code Complexity (demo) ⓘ

Average cyclomatic complexity over time.

Lines of Code (demo) ⓘ

Average lines of code per submission over time.

Early study results

Aggregated findings you get to see as they emerge.

Group trends

See how accuracy and speed are shifting across all participants over time — anonymised and aggregated.

AI use patterns

Early signals on how self-reported AI usage relates to changes in coding confidence and performance.

Live updates

Results are updated as new sessions come in — participants see the study evolve in real time.

Help us answer these key questions

We are looking for participants, researchers, and collaborators.

Start Contributing Collaborator Interest

Support This Research

Please help cover the costs of this app, including servers, storage, and ongoing development. With your support, I'd be able to bring in researchers and developers to add new features and further advance the science.

Get in touch