Receiving an algorithmic recommendation based on documentary filmmaking techniques

TL;DR

This paper explores a novel algorithmic recommendation system for documentary films based on filmmaking techniques, analyzing user reception and proposing improvements for personalized documentary film recommendations.

Contribution

It introduces a metadata framework based on filmmaking dispositifs to diversify recommendation criteria beyond traditional thematic classifications.

Findings

Users engaged with recommendations based on filmmaking dispositifs.

The metadata approach offers a new perspective for documentary film recommendations.

Limitations identified suggest avenues for enhancing recommendation accuracy.

Abstract

This article analyzes the reception of a novel algorithmic recommendation of documentary films by a panel of moviegoers of the T{\"e}nk platform. In order to propose an alternative to recommendations based on a thematic classification, the director or the production period, a set of metadata has been elaborated within the framework of this experimentation in order to characterize the great variety of ``documentary filmmaking dispositifs'' . The goal is to investigate the different ways in which the platform's film lovers appropriate a personalized recommendation of 4 documentaries with similar or similar filmmaking dispositifs. To conclude, the contributions and limits of this proof of concept are discussed in order to sketch out avenues of reflection for improving the instrumented mediation of documentary films.