Machine Learning (ML) For Tracking the Geo-Temporality of a Trend: Documenting the Frequency of the Baseball-Trucker Hat on Social Media and the Runway
Getman, Rachel Rose
The study applied fine-grained Machine Learning (ML) to document the frequency of baseball-trucker hats on social media with images populated from the Matzen et al. (2017) StreetStyle-27k Instagram dataset (2013-2016) and as produced in runway shows for the luxury market with images populated from the Vogue Runway database (2000-2018). The results show a low frequency of baseball-trucker hats on social media from 2013-2016 with little annual fluctuation. The Vogue Runway plots showed that baseball-hats appeared on the runway before 2008 with a slow but steady annual increase from 2008 through 2018 with a spike in 2016 to 2017. The trend is discussed within the context of social, cultural, and economic factors. Although ML requires refinement, its use as a tool to document and analyze increasingly complex trends is promising for scholars. The study shows one implementation of high-level concept recognition to map the geo-temporality of a fashion trend.
trend visualization; Artificial intelligence; Fashion; athleisure; baseball hat; deep-learning; normcore; machine learning
Green, Denise N.
Fiber Science and Apparel Design
M.A., Fiber Science and Apparel Design
Master of Arts
dissertation or thesis