How we work

tanitdata combines rigorous research with practical engineering. Our work follows a small set of principles — the tanitdata Research & Engineering Protocol — designed to keep results useful, transparent, and reusable.

question → evidence → diagnosed mechanism → engineering response → reusable open output

Principles

Start with the question

Each project begins with a clearly defined problem or decision to investigate, not with a dataset or technology looking for a use.

Work from authoritative sources

We prioritize primary and institutional sources and document where the data comes from, how it was accessed, and when it was retrieved.

Assess the data before using it

Accessibility, completeness, consistency, metadata, update frequency, and machine-readability are evaluated as part of the research itself, and published as findings in their own right.

Make the evidence traceable

Findings should be connected to identifiable data, calculations, queries, or records so they can be independently examined.

Make the method reproducible

We document the steps, assumptions, code, and environments needed to reproduce the analysis.

State what the evidence cannot tell us

Uncertainty, missing data, conflicting sources, and methodological limitations are reported alongside the findings.

Engineer from the mechanism

Technology follows the diagnosed mechanism. We use AI, IoT, software, or other engineering approaches only when they provide a credible way to address the problem identified by the research.

Leave something reusable

Substantial projects produce assets others can build upon: datasets, source inventories, code, notebooks, schemas, benchmarks, APIs, MCP tools, prototypes, or documented methods — published openly by default.

Publish at the pace of the question

We release work when it is useful rather than when it is complete, maintain clear versions, and update it as better data, methods, or evidence become available.