Data Analysts

Pre-built reports answer pre-imagined questions. Real analysis requires the underlying data in formats that drop straight into the analyst's stack.

A federation researcher studying youth retention needs license tables joined to results joined to demographic metadata, in a single Parquet file they can pull into pandas in 30 seconds. Without bulk exports, analysts either reverse-engineer the API into a Frankenstein extract pipeline or give up and use whatever fragmentary CSVs the federation has been emailing around.

Both outcomes produce worse insight and slower research cycles.

Bulk data export covers rankings, results, licenses, demographics, events, and clubs in CSV, Parquet, and newline-delimited JSON. Schemas are documented with field types, allowed values, and lineage notes per column.

Exports are versioned and dated; re-running the same export with the same parameters yields identical bytes for reproducibility, and historical export versions remain downloadable for replication studies.

Academic publication and federation policymaking both rest on data the analyst can defend under questioning. That means stable identifiers across exports (a player ID this year is the same as last year), explicit handling of merges and deduplications, and per-record provenance — what source contributed this row, when, and through what validation.

Without that backbone, peer reviewers reject the work and policymakers distrust the conclusions. Quality is what turns export volume into publishable research.

Every dataset carries stable platform identifiers with documented merge and deduplication rules applied transparently. Per-record provenance metadata captures source system, ingest timestamp, and validation status.

Validation rules and known data-quality issues are published per release so analysts can apply appropriate exclusions and document them in their methodology with confidence under peer review.

Many analytical questions don't justify a full export. An analyst asking 'what's the median age of competitive female players in regions X, Y, Z over the past three years' needs an answer in seconds, not a workflow that downloads 800 MB of license data.

A custom query interface — SQL-style or GraphQL — gives analysts ad-hoc aggregation power, lets them iterate quickly, and keeps sensitive raw data on the platform when only the aggregate is needed. It also enables exploratory work that bulk exports discourage because of their friction.

The custom query interface exposes a GraphQL surface for relational traversal and a read-only SQL endpoint for aggregations against curated analytical views. Both honor the analyst's data-access scope, log every query for audit, and return results streaming where possible to support large aggregations without timeouts.

Saved queries can be shared across research teams with version pinning.