Passer au contenu principal
Accéder au tableau de bord
Vous ne savez pas par où commencer ? Répondez à un bref questionnaire pour obtenir des recommandations personnalisées.
NZME

NZME enhances data quality in the newsroom

NZME adopts Dataform, enhancing data model reliability and quality.

The Challenge

In the fast-paced and ever-evolving landscape of the modern newsroom, journalists face an array of time challenges. With the 24/7 news cycle and the need to publish and understand performance across a variety of channels, reporters need high quality and timely data at their fingertips.

An in-depth review of systems and processes concluded that although the newsroom had access to a wide array of reports and dashboards, these often surfaced duplicated or conflicting information. NZME needed to rebuild their data models using best practice tools and processes to ensure the data delivered was accurate and trusted, and then deliver a consolidated suite of editorial reports to enable reporters to find the information they needed quickly and efficiently.

NZME - Newsroom Data Transformation_Headshot_andy_wylie_photo
"NZME were early adopters of Dataform and have been working with Google since the product was in Beta. By implementing best practices acquired during an extensive workshop at Google's offices, we enhanced our Dataform workflows and are now at the forefront of Dataform implementation for publishers. The adoption of Dataform delivered a step change in the explainability, reliability and quality of our data modelling pipelines, simultaneously enhancing efficiency and reducing technical debt."
Andy Wylie
Head of Data and Analytics, NZME

The Results

To deliver the insights needed by the newsroom, the NZME analytics team chose to work with granular, event level data streamed directly into the Google Cloud BigQuery Data Warehouse. Data quality is paramount, especially when wrangling millions of rows daily. To establish a robust foundation for the data models, the team looked to Google Cloud’s Dataform.

Dataform is a data transformation and orchestration tool natively integrated into BigQuery making it an obvious choice for NZME’s analytics team. Dataform’s integrations with code repositories and automated documentation helped analysts collaborate more efficiently, reducing the overhead of building and maintaining data models. Additionally, the modular structures built within Dataform encourages code reusability, driving simplification of data pipelines and ultimately a reduction in query cost.

Dataform has provided NZME with dramatically improved visibility into their data models and pipelines. Analysts can get up to speed more quickly on new datasets, and easily understand upstream dependencies and downstream impacts by reviewing the Dataform compiled graph. The code repository integration means that changes are documented and peer reviewed before deployment, and can be rolled back easily if needed. Dataform "assertions" have enabled NZME analysts to quickly and easily build data quality checks into the pipelines that trigger alerts if issues are detected.

The adoption of Dataform delivered a step-change in the explainability, reliability and quality of NZME reporting and formed the foundation for a new suite of editorial reports and dashboards, ensuring that the newsroom has the information they need at their fingertips.

  • 0% downtime of the Dataform platform
  • >250 production tables and views updating daily via Dataform pipelines
  • 50% decrease in time taken to train newly-hired analysts on existing datasets and pipelines orchestrated via Dataform
NZME - Newsroom Data Transformation_Image 1
Dataform graph view of simple data pipelines
NZME - Newsroom Data Transformation_Image 2
Example of an NZME newsroom dashboard showing 4 bar and line charts.
NZME - Newsroom Data Transformation_Image 1
Dataform graph view of simple data pipelines
NZME - Newsroom Data Transformation_Image 2
Example of an NZME newsroom dashboard showing 4 bar and line charts.
Quitter et perdre la progression ?
Si vous quittez cette page, vous perdrez votre progression dans la leçon en cours. Voulez-vous vraiment continuer et perdre votre progression ?