The Eagle platform allows data standardization according to common models, usable in all countries where Carrefour is present. The platform also develops data models addressing local market's needs, in order to build data science analyses and integrates the possibility to implement artificial intelligence algorithms. By analysing the data from the platform, the retailer can now observe shopping patterns or can segment clients, information that will enable a data-driven decision-making process.
With over 10,100 stores in 34 countries, the Carrefour Group is the world's second-largest retailer and number one in Europe. Every day, over 10 million customers visit Carrefour stores around the world, enjoying a wide range of products and services at affordable prices. In Romania, the Carrefour group offers its customers multiple shopping possibilities, both in stores or online, through the unique portal www.carrefour.ro or the BRINGO delivery service.
The main technical difficulties that had to be overcome derived from the very large volumes of data, the need to process them in a unitary format, standard at the Carrefour group level. All these challenges led to considerable difficulties in data analysis and in implementing a data-driven decision-making process. Another challenge was the development of a replicable solution for all countries in which the retailer is present, which would allow easy implementation of new machine learning and AI algorithms. Therefore, the platform represents the foundation for implementing algorithms that substantiate the pricing, loyalty, customization, or supply policies.
Eagle, Carrefour's big data platform on Google Cloud, centralizes daily information from the retailer's main data sources, leading to a data-driven vision of the company. There are millions of records that are processed and managed every day. All the aggregated data is suited to be processed through AI algorithms. Therefore, the process will highlight insights about interaction patterns, and customer segmentation, enhancing the efficiency of promotional campaigns In the following stages of the project, new data sources, machine learning, and AI algorithms will be implemented, for actual data processing.
The project was implemented in just 6 months by a team of big data experts from eSolutions, including three Google Cloud Certified Platform Architects, a Google Cloud Certified Professional Data Engineer, a Big Data Developer, and a Project Manager.
Google Cloud Platform tools (Pub/Sub, Cloud Functions, Airflow, BigQuery, GCP Monitoring), Terraform, Terragrunt, Scala, Apache Beam, Scio