By Xavier Gumara Rigol, Senior Engineering Manager for Experimentation and Analytics Solutions and Sandra Real, Product Manager for Analytics Solution

Building a data mesh to support an ecosystem of data products at Adevinta

submited by

Style Pass

2021-06-18 13:30:03

By Xavier Gumara Rigol, Senior Engineering Manager for Experimentation and Analytics Solutions and Sandra Real, Product Manager for Analytics Solutions at Adevinta.

In this article, we’ll explain our journey towards a data mesh architecture. It began in 2018 (prior to the data mesh concept), when our Analytics Solutions team received the mandate to enable convenient access to our marketplaces’ data…

The Central Product and Tech department (P&T) in Adevinta provides a wide range of data products that our marketplaces can leverage (amongst other services). Some examples include an ads moderation service, recommender systems, image recognition algorithms and chat services.

In order to measure the impact these products have on our end-users and make the best possible decisions, we use data. It is therefore crucial to capture, store, clean and make data easily available to decision makers in the company: Data Analysts, Data Scientists and Product Managers.

To do so, we’ve embraced the concept of a data mesh as proposed by Zhamak Dehghani: a federated model for self-service management of the data from our Central P&T components.