Unsupervised Behaviour Analysis of News Consumption in Turkish Media

TL;DR

This paper analyzes anonymized clickstream data from Turkish news websites to identify consistent news consumption patterns over time using ensemble clustering and advanced embedding techniques, providing insights for media organizations.

Contribution

It introduces a novel ensemble clustering approach with mixed-type embeddings and validation strategies to analyze news consumption behaviors from clickstream data.

Findings

Clustering with mixed-type data yields high internal validation scores.

UMAP embedding combined with ensemble methods improves cluster quality.

Identified stable reader groups over monthly periods with high similarity.

Abstract

Clickstream data, which come with a massive volume generated by human activities on websites, have become a prominent feature for identifying readers' characteristics by newsrooms after the digitization of news outlets. Although the nature of clickstream data has a similar logic within websites, it has inherent limitations in recognizing human behaviours when looking from a broad perspective, which brings the need to limit the problem in niche areas. This study investigates the anonymized readers' click activities on the organizations' websites to identify news consumption patterns following referrals from Twitter,who incidentally reach but propensity is mainly routed news content. Methodologies for ensemble cluster analysis with mixed-type embedding strategies are applied and compared to find similar reader groups and interests independent of time. Various internal validation perspectives are used to determine the optimality of the quality of clusters, where the Calinski Harabasz Index (CHI) is found to give a generalizable result. Our findings demonstrate that clustering a mixed-type dataset approaches the optimal internal validation scores, which we define to discriminate the clusters and algorithms considering applied strategies when embedded by Uniform Manifold Approximation and Projection (UMAP) and using a consensus function as a key to access the most applicable hyperparameter configurations in the given ensemble rather than using consensus function results directly. Evaluation of the resulting clusters highlights specific clusters repeatedly present in the separated monthly samples by Adjusted Mutual Information scores greater than 0.5, which provide insights to the news organizations and overcome the degradation of the modeling behaviours due to the change in the interest over time.