Photos are a wonderful medium in corporate and science communications to showcase what faculty, staff, and students do at the ZHAW. But which photos are interesting and which are not? The aim of our project is to better understand the aesthetic and semantic characteristics of the photos in relation to the academic context: what types of images are viewed, liked, and shared on social media and photo-sharing platforms and what makes them appealing? To answer this question, we will combine machine learning with methods from visual semiotics. We will perform a data-driven analysis to spot patterns in the photos by extracting structural, holistic scene features, and semantic information of the photos. The findings are intended to enhance the digital communications strategy of the ZHAW on social media. The research project is a collaboration between Media Linguistics (visuals semiotics) and Computer Vision (machine learning/deep learning).