← Back to case studies
AI-Native Apps

How a University Course Designer Turned Three Earlier Dashboard Concepts Into a Live, 100-Student Trade-Policy Simulation in Six Days

A fast, test-first build for a fixed live-event date, with bugs caught before the students arrived, not during the event.

Client snapshot

University College Dublin runs ECON30160, a trade-policy course that uses a live, in-person negotiation game: students split into small groups at physical tables, each representing a geopolitical role (China, Russia, the USA, the EU, India, Africa, South America), working with paper cards and a facilitator. Marc Leberecht-Schneider designed the game and needed a digital dashboard to run alongside it.

The situation before

By the time this project started, Marc had already built and tried three separate dashboard designs, each a different way of visualising the game's state for facilitators and students. None had been taken forward. The live event date, 2 November 2026, was fixed: around 100 students, playing across several tables at once, each representing up to seven groups.

Why they acted, and why now

Marc and Andreas had already worked together on other web projects. With a hard deadline approaching and three prior dashboard attempts behind him, Marc brought the raw idea back to Andreas rather than starting a fourth attempt alone or from scratch with someone new.

Goals & success criteria

  1. 1A dashboard reliable enough for roughly 100 students to use without facilitator help.
  2. 2Built fast enough to be ready for 2 November.
  3. 3Aligned closely enough with the physical card-and-paper game that both halves felt like one experience.

What we did

Work started on 27 October 2026, six days before the live event. Instead of tackling the full, complex idea at once, the two worked in short cycles: strip the idea down, build a small piece, test it, get feedback, repeat. Andreas built the simulation engine (21 real intervention cards, five wellbeing dimensions, a four-camp mindset system, external-shock events, round-based weighting and voting) and the dashboard UI on top of it, backed from early on by unit tests and an end-to-end test suite run before every change shipped.

About a week before the live event, Marc ran his own dress-rehearsal gameplay test. It surfaced a list of weak points in the game logic and interface. That list became the next round of work, handled in small, tested pieces rather than as one large rewrite. Two real bugs came up this way: a rounding mismatch in the camp-mindset calculations, caught by a later code review, and a comparison-view bug where two of three scoring views wrongly showed identical numbers for a camp that hadn't yet entered its own data, which Marc flagged and Andreas fixed the same day. Both were caught before the live event, not during it.

Results

How we worked together

Marc supplied the ground-truth data throughout: card fiscal values, mindset weightings, and written diagnoses of bugs he found. Andreas worked those into the codebase in small, tested steps, kept every number traceable back to Marc's own source files, and ran a code review before every change that reached the live site.

Why this worked

This way of working depends on being comfortable with an idea that starts rough and improves in small steps, rather than needing it finished and polished from the first version. For a project with a fixed date and real users at the end of it, catching problems early, through a dress rehearsal and disciplined testing, mattered more than any single feature being perfect on the first try.

Who this is for

Course designers, workshop facilitators, or anyone running a dated live event who already has a rough concept, or has tried earlier attempts that did not go forward, and needs a working tool fast. It is not the right fit for someone who expects a finished product on the first pass, or who is not willing to test and adjust along the way.

Frequently asked questions

Did the dashboard replace the physical card game?

No, it ran alongside it. Students still worked with paper cards at physical tables, and the dashboard tracked and displayed the results.

How long did this take to build?

Six days from the first commit to the live event, with ongoing refinement afterwards.

Was any student data stored?

No. The dashboard has no backend; all state lives only in the browser during the session.

If you're running a dated event with a rough concept, or earlier attempts that didn't go anywhere, let's talk.

Prefer email? andreas@mindfulai.academy