PREDICTING VISUAL AESTHETIC PREFERENCES OF LANDSCAPES NEAR HISTORICAL SITES BY FLUENCY THEORY USING SOCIAL MEDIA DATA AND GIS

There is an interactive relationship between humans and landscapes. Humans inherently assess landscapes by creating spontaneous preferences based on surrounding stimuli. Vision plays a key role in these preferences. Visual preferences are relevant for understanding visual aesthetic liking (VAL), which needs to be evaluated objectively. This study was carried out in Herakleia ad Latmos, comprising Lake Bafa Natural Park and the Latmos-Besparmak Mountains. The aim of this paper is to predict people’s VAL of historical sites (HS) by applying processing fluency theory to social media data. Among fluency theory metrics, four metrics – visual simplicity, visual symmetry, visual contrast, and visual self-similarity, were used to develop an ordinary least squares (OLS) regression model. Two primary questions are explored in this study: (1) How to quantify spontaneous visits of people near historical sites, and (2) how to estimate preferences of people based on distances to HS regardless of landscape types (either cultural or natural). Results show that people mostly visited three HS out of thirteen historical sites between 2004 and 2020: Kapikiri Island (HS 1), and the ancient cities of Herakleia (HS 2) and Latmos (HS 3). According to the findings of the OLS regression model, year (t = 8.99, p <.0001), visual simplicity (t = -4.64, p ≤ 0.0001), and visual contrast (t = -2.01, p = 0.04) of the geotagged photos were all statistically significant predictors of VAL. HS 2 had the highest VAL value, followed by HS 1, and HS 3.