Context and problematic

The Data team of this large French media company is in charge of analyzing the traffic and actions taken on their website,
In a context where content personalization is more and more present on websites, it is also in charge of the personalization of the website, thanks to the large amount of data generated every day.

Goals

Accompany the client in the reflection and implementation of methods to personalize the content for the users. The objective is to build reader loyalty and increase visits to the site.
Visualization and analysis of the traffic on the site.
Creation of a personalized daily newsletter

Our intervention

1 Data Scientist, 1 Data Analyst

• Definition of KPIs to monitor traffic trends on the site.
• Improved score based on article reading, alert activation and site navigation.
• Creation of a personalized newsletter based on this score and business rules.
• Automation of the creation process of this personalized daily newsletter.
• Etat de l’art des méthodes de recommandation pour pouvoir recommander 10 articles récents aux lecteurs.
• State of the art of recommendation methods to recommend 10 recent articles to readers.
• Model performance comparison
• Putting the chosen solution into production

Results

Creation of DashBoard to visualize the KPIs, followed by the sales teams and journalists.

Validation of the results of the personalized newsletter by the sales team and implementation of an A/B test.

Technical environment

Google Cloud Plateform, BigQuery, SQL
Data Studio, Qlik Sense
Python, TensorFlow, pandas, Jupyter