Portfolio
Work in Progress · In ArbeitEvidoxa — Research Data Management for Humanities Scholars
Historians work daily with complex networks of people, places, sources and events — but hardly any tool is built for them. Existing solutions are effectively designed for IT departments: powerful, but steep in learning curve, demanding in prerequisite knowledge, and built around interfaces that intimidate rather than invite. Many humanities scholars aren't even aware these tools exist.
Process
Phase 01
Problem Discovery
Conversations with historians revealed a recurring need: a single place to bring together and meaningfully analyze historical data, locations, people, sources and relationships — without requiring technical expertise.
Phase 02
Vibe Coding as a Design Method
No wireframes, just code. Using Cursor AI, Next.js and MUI, usable interfaces emerged quickly — prototyping not as a precursor to development, but as a design tool in its own right. What works, stays. What doesn't, goes.
Phase 03
Database Architecture as a UX Problem
It became clear early on that information architecture cannot be translated 1:1 into database structures. An intensive learning journey into data modeling followed — with the goal of building a structure that offers maximum flexibility and scalability without passing that complexity on to users. The data model was rebuilt from the ground up multiple times.
Phase 04
Scope Expansion Through Feedback
Historians rarely research alone. Based on user feedback, collaborative projects were added alongside integration with bibliography platforms for referencing secondary literature.
Phase 05
POC Preparation
The next step: broader validation with a larger group of researchers — less a classic usability test, more a proof of concept in a real research context.
Methods
AI-aided Development
Cursor AI as a thinking partner and development tool — not for automation, but to accelerate learning processes and design decisions.
Database Modeling
Iteratively rebuilding the database structure as an integral part of the UX process — because the backend defines the limits of the frontend.
Code as Prototype
No Figma, no Balsamiq — prototyping directly in production-ready code to make real interaction tangible from the start.
User-centered Scope Decisions
Features driven not by ideas, but by conversations with the target audience — and the willingness to expand scope when feedback demands it.
Visuals
Main interface / Dashboard
Complex data structure, simply presented
Key Takeaways
Prototyping with Code
The most direct path from problem to testable solution — if you're willing to embrace the learning curve.
Complexity is a UX Decision
A powerful data structure in the background doesn't demand a complex interface in the foreground. Designing one without understanding the other doesn't work.
Need ≠ Solution
A technically solved need is worthless if the solution was built past the mental models and daily reality of its users.
Modularity as a Survival Strategy
Scope changes. Building for that from day one means surviving the changes. Not building for it means building twice.
Evidoxa is in active development. The current state is a functional MVP with user management, collaborative projects, a flexible data model and bibliography integration. The next step is validation in a real research context.