Plainsight

Ads & Data · Media & entertainment

Every media type runs on its own data architecture. How do we get one view across all of them?

Plainsight modernised the data platform of Ads & Data, replacing a mix of legacy architectures, one per media type, with a single platform on Azure Data Lake, Databricks, Airflow and Power BI that serves more than 1,000 report views a day.

  • Display, video, print and TV
  • Azure Data Lake, Databricks, Airflow, Power BI
  • 1,000+ report views a day
  • 100+ report users a day

Uitdaging

A different architecture for every medium

Ads & Data works across very different media: display and video, print, and TV advertising. Over time, each media type had grown its own data architecture, built around its own sources and its own logic.

That works until you want one view across all of them. Then every comparison turns into a reconciliation exercise, and every change has to be made more than once.

Wat we deden

One platform instead of many

One home for every medium.

We brought the media types together on one modern data platform: Azure Data Lake for storage, Databricks for processing, Airflow to orchestrate the pipelines and Power BI for reporting.

Our Databricks practice →

Built by the book.

We built it to best practice throughout. A platform this size only stays maintainable if it follows clear conventions, so the next person who works on it can find their way around.

Pipelines that run on schedule.

Airflow decides what runs when and in which order, so data lands in the right sequence and a failed run is visible instead of silent.

Reports on one version of the truth.

Power BI sits on top of the shared platform, so a report on TV and a report on display start from the same data and the same definitions.

How we build data platforms and analytics →

Impact

A platform people use

One view across display, video, print and TV, instead of one per medium.

  • Up to 5x faster time to insight
  • More than 1,000 report views a day
  • More than 100 unique report users a day
Data platform modernisationAzure Data LakeDatabricksAirflowPower BI

Veelgestelde vragen

Why bring separate data architectures together?

Because the questions the business asks cross media types. One platform means one definition of each number, one place to fix it, and comparisons that don't need reconciling first.

Why use Databricks and Airflow together?

Databricks does the heavy data processing; Airflow decides what runs when and in which order. Keeping orchestration separate makes the pipelines easier to monitor and easier to extend with new sources.

How does a data platform cope with large data volumes?

By keeping storage and compute separate. The data sits in the lake at low cost, and Databricks scales compute up for the heavy jobs and back down when they are done.

Een gelijkaardig probleem bij jou?

Contacteer een expert