Ghost Data Tool (Prototype)
In most data workflows, analysis begins with cleaning and modeling, often skipping the step of questioning who the data represents, and perhaps more importantly, who it doesn’t. Ghost Data emerged from my desire to build in a moment of critical reflection before technical work begins.
The "Ghost Data" tool is a an AI-trained app that generates personalized critical reflection questions about who is excluded and who benefits in a given dataset. It analyzes datasets against the seven principles of data feminism, according to Catherine D'Ignazio and Lauren Klein. Designed for data scientists, creatives, journalists, academics and activists (anyone working with data), Ghost Data will help its users reverse engineer their dataset to identify biases in their field and imagine how today’s exclusions in data could shape future injustices.
Creative Strategy + Execution
Built Figma mockups for a user flow where datasets can be uploaded and analyzed against those principles.
Leveraged “vibe-coding” and Cursor to prototype functionality, with backend API support from collaborators to connect the system to an AI engine.
The tool generates custom reflection prompts — e.g., how the dataset challenges power, whether it perpetuates binaries, or whose labour is made invisible, to spark team discussions before any modeling or statistical evaluation.
Results
Still in its prototyping phase, Ghost Data functions as both a design probe and a practical tool. It represents an experiment in merging UX thinking with responsible data science practices, making reflection an intentional step in the workflow.
