The Problem

Large companies rarely stick to one BI tool. Finance works in Power BI, marketing in Tableau, and an older team is still on Qlik. Over the years that turns into thousands of dashboards nobody can fully account for. It's a bit like streaming subscriptions: you know you pay for several, but you're not sure which ones anyone still watches.

Each BI tool can report on itself. None of them can answer the questions that sit between them: what do we actually have, who is using it, and are we paying for licenses nobody touches?

What I Built

BI360 is the cross platform analytics view inside Datalogz Control Tower. I built it end to end during my time as a founding engineer at Datalogz, from the data pipelines all the way to the screens executives use. I also wrote the official product post about it, which is still live on the Datalogz blog: Getting a 360° View of Your Analytics Estate.

It gives BI admins and data leaders three things:

• Global Inventory: one searchable catalog of every workspace, dashboard, report, and dataset across all connected tools.
• Usage and health: who viewed what and when, which assets have gone stale, and which data refreshes are failing before anyone notices.
• Organizational adoption: usage mapped to real business units through Azure Active Directory, so leaders can see which teams use their licenses and which don't.

The payoff is practical. Teams retire dashboards nobody opens, move expensive licenses to people who need them, and know the business is running on healthy, maintained data.

How It Works

I treated BI metadata like any other data engineering problem: extract it, normalize it, serve it.

1. Extraction: daily Apache Airflow jobs call the Power BI, Tableau, Qlik, and Spotfire APIs and pull users, permissions, workspace structure, and view counts into the data warehouse.
2. Normalization: the hardest part. A Power BI workspace and a Tableau project mean roughly the same thing, but they look nothing alike in raw data. I wrote dbt models that map every platform into one shared BI360 schema, so the rest of the system only ever sees a standard asset, no matter which tool it came from.
3. API: a FastAPI service sits on top of the warehouse. Every request is checked against the user's permissions, scoped by filters like department, and then run against the normalized data.
4. Interface: a React frontend turns all of that into the product, with data grids that sort and filter thousands of assets and charts that plot Tableau and Power BI usage on the same axis.

The Design Choice That Mattered

The normalization layer is the decision that made BI360 work. Because every tool lands in one schema, supporting a new platform means one new connector and one new dbt mapping. The API and the interface don't need to change.

Read the Full Story

For the product side, including the executive views and the governance problems BI360 solves, read the original post I wrote on the Datalogz blog: Getting a 360° View of Your Analytics Estate.