How to connect the API to BI dashboards (Power BI, Looker, Tableau)

Power BI, Looker, Looker Studio and Tableau are where most teams want Tracksuit data to live: building scorecards, brand-health dashboards, or blending with sales and media spend. None of them connects to the Tracksuit API the way you might expect. This guide covers the realistic option for each, and what to get right once the data lands in your BI layer.

Whatever your BI tool, we recommend loading Tracksuit data into a data warehouse first, then pointing your BI tool at that. While some tools can technically call the API directly, the warehouse is the only path that stays reliable once reports are scheduled and shared.

👉 Set up How to connect the API to your data warehouse before you start.

👉 Try it in the browser first. Our interactive API documentation lets you paste in your key and run live requests, so you can see the exact response shape before wiring up a dashboard.

Why we recommend a warehouse for every BI tool

The Tracksuit API is a read-only REST API with four traits that BI tools handle poorly out of the box:

Bearer-token auth. Every request needs an Authorization: Bearer YOUR_API_KEY header. Easy in code; awkward in a BI tool's built-in web connector, and a blocker for unattended scheduled refresh.
Cursor pagination. List responses return a next_token; you must loop until it comes back null. Connectors don't do this automatically. Stop early and you silently load partial data.
JSON, not tables. Responses are nested JSON that needs flattening before it's chart-ready.
Rate limits. 5 requests per second (burst 10). A warehouse pulls once on a schedule; a BI tool refreshing per-user can trip 429s.

A warehouse solves all four once, for every downstream tool. That's why it's the recommended foundation rather than a workaround.

Decision guide

Tool	Direct connection supported?	Recommended path
Looker (LookML / enterprise)	❌ No. Connects to SQL databases only, so it cannot call a REST API	Warehouse, then Looker
Looker Studio (free, formerly Data Studio)	⚠️ Only via a community connector	Warehouse (BigQuery), then Looker Studio
Tableau	⚠️ Refresh limitations. Not a supported or reliable setup	Warehouse, then Tableau
Power BI	⚠️ Refresh limitations. Not a supported or reliable setup	Warehouse, then Power BI

Recommended: connect through a warehouse

This is the robust path for all four tools. Load the API into Snowflake, BigQuery, Redshift or Databricks once (see How to connect the API to your data warehouse), then connect your BI tool with its native database connector, with no API code in the BI layer.

Power BI: Get Data, then your warehouse, then Import or DirectQuery.
Looker: point a LookML connection at the warehouse and build one explore per fact table (funnel, conversion, statements), each joined to the category_views dimension.
Looker Studio: add your warehouse (BigQuery works natively) as a data source.
Tableau: connect to your warehouse with the native connector and build off the typed tables.

Because the warehouse holds clean, typed tables (one per endpoint, keyed by category_view_id), every tool gets fast, refreshable, governed access, and pagination, auth and rate limits are handled upstream where they belong.

"Looker" vs "Looker Studio". They're different products. Looker (LookML) is SQL-only — warehouse required. Looker Studio (the free, formerly Data Studio tool) can connect to BigQuery directly, so if you land Tracksuit data in BigQuery it works with no extra connector; otherwise it needs a community connector.

Don't connect Power BI straight to the API for scheduled or shared reports. Automatic refresh won't work without extra setup, repeated refreshes hit the request limit, and data can load half-finished.

Get these right in the BI layer

However you connect, these Tracksuit-specific rules apply when you build the visuals:

percentage is a 0–1 fraction. 0.42 means 42%. Format it as a percentage in the visual, don't multiply by 100 in your model, or you'll double-count when someone formats it again.
Don't sum demographic slices. Weighting is applied dynamically per slice, so age groups won't add up to the total. Let users pick a slice or the unfiltered total. See How to filter by demographics.
Data moves in monthly waves. Refresh monthly (or weekly to be safe), not hourly. There's no new data between waves. See How data freshness & wave timing work.
Mind the rate limit on direct connections. 5 requests/second, burst 10. A per-user dashboard refreshing straight against the API can hit 429s; a warehouse pulls once and avoids the problem entirely. See Tracksuit API rate limits.
Tokens expire after 365 days with no refresh mechanism. Set a calendar reminder to rotate the key before it lapses, or your dashboards will quietly stop refreshing.