Automated Semiotic Analysis: Hero, Heretic or Helper?

NShapes sortedewMR Webinar – 6 July 2022

Click here to access the slides and recordings.

Andrew Jeavons, co-founder Signoi

Automated Semiotic Analysis: Hero, Heretic or Helper?

You will learn:
How to make sense of large amounts of images, text and other unstructured information – blending AI with humans to handle bigger datasets.

Click here to access the slides and recordings.

Traditionally semiotics has been a ‘practitioner’ based analytic method. An individual, or perhaps a group, performs an analysis of text, images, transcripts et al. There may be some cross checking of interpretations in some cases.
The volume of material that can be used in this analysis is limited to the capacity of the individual or team. People have finite amounts of material they can absorb and analyse and this is a fundamental limitation of the practitioner approach within semiotics in the age of big data. The volume of images that can be processed is perhaps hundreds, text has similar limitations imposed by time scales of projects.

In this talk we present the results of automated semiotic analysis of thousands of images and many thousands of lines of text/tweets. This type of approach is seen by some as an anathema to the semiotic method. However, there are, we believe, cogent arguments for the use of this approach. Not least of these arguments is the massive volume of social media posts (Instagram, Twitter and many more) that are a significant part of a brands ecosystem and need to be analysed at scale. Semiotics is a powerful approach to the analysis of this type of data and it runs the risk of being marginalized because of the volumes of information that are being generated. By leveraging techniques from machine learning we believe automated analysis of large datasets is a necessity for the future of semiotics. This includes not only AI-driven thematic analysis to surface natural semiotic structures, but also metrics such as Archetype analysis. It should be seen as an aid to data reduction in order to facilitate deeper analysis by the expert practitioner.