What is the Color of a City? Extracting Color Representations Using Flickr Images

dc.contributor.advisorGiordano, Alberto
dc.contributor.authorShabalina, Elena
dc.contributor.committeeMemberChow, Edwin
dc.contributor.committeeMemberBlue, Sarah
dc.description.abstractThe world we live in is inextricably linked to the internet which is becoming a full-fledged replacement of classical tools for obtaining information and communication. People often take decisions after browsing the internet. For instance, prior to traveling, people often look at pictures before deciding where to go on vacation and make their judgement based on the place’s attractiveness. This can lead to problems such as over-tourism (Milano et al., 2018). Additionally, we are influenced not only by the content, but also by images’ colors (Madden et al., 2000). This study aims to determine the relationship between cities and color by analyzing images uploaded to Flickr. Specifically, it investigates color variability in general, finds relationships between color palettes and content for ten selected cities, and looks for color differences between cities. To test the hypothesis that different world cities are similar to each other based on colors people see on Flickr, I proposed an adaptive algorithm for extracting main colors from images using grayscale conversion. I created an interactive map of Texas using random geolocated Flickr images. To find the relationship between color and content, as well as the geographical distribution of color, MANOVA and other statistical analyses were conducted. The results show that each city has a unique color palette characterized by content and sociocultural impacts. Also, there are territories inside cities with a unique color that describe these districts in their own way and may not correspond to the perception of color that we see in real life. In addition, the color variation among cities is not geographically random. Specifically, cluster of greens confidently contains the same content (vegetation), while other colors, such as blue and gray, contain unique information about real-world objects (~ 60% for blue, ~ 62% for gray). The results suggest that nature is more uniform in color, while the life of people and their social interactions have many bright colors and, as a result, different content. It is possible that old cities, which usually have very distinct architecture, differ from newer cities, where social events take precedence over historical content. On this basis, the concept of color on social media images can be used in digital and urban geography.
dc.description.departmentGeography and Environmental Studies
dc.format.extent79 pages
dc.format.medium1 file (.pdf)
dc.identifier.citationShabalina, E. (2021). What is the color of a city? Extracting color representations using Flickr images (Unpublished thesis). Texas State University, San Marcos, Texas.
dc.subjectsocial media
dc.subjectcolor analysis
dc.titleWhat is the Color of a City? Extracting Color Representations Using Flickr Images
thesis.degree.departmentGeography and Environmental Studies
thesis.degree.grantorTexas State University
thesis.degree.nameMaster of Science


Original bundle

Now showing 1 - 1 of 1
Thumbnail Image
3.79 MB
Adobe Portable Document Format

License bundle

Now showing 1 - 1 of 1
No Thumbnail Available
2.56 KB
Item-specific license agreed upon to submission