Fashion is Hidden in the Big Data of Instagram
What will be in vogue next season? Paola Cillo, Associate Professor of Management and Technology, has asked an algorithm. Prediction in times of uncertainty is one of the research areas on innovation in the creative industries. This gives rise to the idea of using big data to pinpoint forthcoming successful fashion trends. The research, which is still ongoing, is based on pictures posted on Instagram, but it could be based also on images provided by public cameras. The photo sharing service, that has recently announced that its community has grown to more than 500 million users, is the social media favoured by fashion companies and consumers. "You can identify emerging trends by analyzing photos posted online", Cillo says.
A computer vision software scans every dress image. An algorithm sorts them according to five variables: length, style, graphic, color, material. "The most recurring combination of these variables identifies a dress 'statistically in vogue'. You can measure the distance from the fashion of the day by comparing other images to the above-mentioned 'average dress'. When the differences exceed a threshold, the algorithm identifies an innovation. The latter becomes a trend when it is detectable in a large number of dresses". It would be prohibitive to analyze every picture posted on Instagram. So, the study focuses on a number of trendsetting cities. An algorithm identifies them by screening historical data using geolocation. "Algorithms offer an innovative contribution to scientific prediction", Cillo says. "This process could be useful to the fashion companies creating seasonal collections and to retail buyers".
Read the article by Emanuele Borgonovo on our lives with algorithms and how they enter the work of Bocconi researchers in various fields
Emanuele Borgonovo. Life is but an Algorithm ...
Guido Alfani. New Eyes to Read History
Valentina Bosetti. Researching Climate and Policy with WITCH
Claudia Imperatore. How to Tell if Your Budget Has Been Manipulated
Silvio Petriconi. Studying Banks with Python
Sonia Petrone. Between Theory and Uncertainty: Interpreting the World
Oreste Pollicino. The Law of Privacy and Business Formulas
Gaia Rubera. The Perfect Startup Pitch? An Algorithm Can Write the Model