About tanitdata

Identity

A studio based in Tunis, working across North Africa and the Global South.

Public data should reach the people whose decisions it shapes. Today it rarely does — the data is fragmented, the analysis lives in formats few people read, and the audiences who need evidence most are the ones served least. tanitdata is designed for the interfaces that now structure information work: AI-native access to open data, analysis grounded in that data, and publishing designed for how audiences actually read. Access, analyze, inform — three stages of a single practice, demonstrated through our work and the research we enable.

Access, analyze, inform

Access. Open data is of no use to audiences who can’t reach it in the interfaces they actually use. We build the infrastructure that makes open data directly addressable by AI assistants, so that reaching a dataset is as immediate as asking a question.

Analyze. Infrastructure produces no meaning on its own. The studio and its research fellows turn accessible data into conclusions that are specific, contestable, and grounded in the empirical record.

Inform. Analysis that doesn’t reach its audience doesn’t shape decisions. We publish in formats designed for how public actors, funders, journalists, and civil society actually read — short, visual, shareable — so that evidence reaches the people whose decisions it should inform.

A distinctive voice runs across the three stages: our public positioning on how open data should be made accessible in the AI era, with particular attention to the Global South, whose institutions cannot simply import solutions designed elsewhere.

Principles

• AI-native. Infrastructure, research, and analysis designed for the age of AI assistants — not retrofitted.
• Autonomy. We are no company’s subsidiary. Partners contribute to projects, not to the studio’s governance.
• Empirical grounding. Strategic analysis rests on concrete technical work. The inverse produces empty recommendations.
• Public good. Open data must remain openly accessible — including through AI interfaces.
• Transparency. Methods, code, and data are publicly versioned where context allows.