A first day on the job, built for practice
A simulated company where data learners work through an incomplete brief, ask colleagues for context and use real code to explain their findings.
Unreleased · In development
My role. Conceived and designed the experience; building the product end to end.
Inside the prototype
- Apps in the revenue-case desktop
- 7
- Simulated colleagues to consult
- 5
- Review score to work towards
- 80+
Finding the question is part of the job
A textbook exercise supplies a dataset and a question. At work, the useful context might be spread across a meeting, an old memo and colleagues who each know one piece of the answer.
I built this experience for Data Science and Data Analytics learners to practise that missing part of the job. They arrive as a new analyst at a simulated company, with a work laptop, documents, colleagues and a deadline. Deciding where to start is part of the exercise.
One company, seven apps, an incomplete brief
In the fictional revenue case, the company has missed its quarterly target by $1.9M. The learner must deliver a one-page memo explaining the three biggest causes in dollars.
Briefing contains the assignment, meeting recording, transcript and scoring rubric. Wiki holds data definitions and company notes. Chat connects the learner to five colleagues, each with different context. Sheets offers quick views of the data; BigQuery provides a SQL console; Notebook runs the analysis. Terminal supplies a command-line shell within the desktop.
The clues are deliberately split across these places. A warning in the data dictionary or a conversation with the right colleague can change the analysis before the learner starts coding.
Real Python, followed by a review
The notebook runs Python in the browser through Pyodide, with the dataset already loaded. Learners can test their own approaches without installing a coding environment.
The rubric is visible from the start. Submitting the work returns feedback through chat, and learners revise towards a score of 80 or more. The loop makes reading, asking and improving part of the work.
The prototype is still in development. The figures above describe its design; learning impact has not yet been measured.