My project was a topic map of 349 photographs taken by photographer Charles Cushman from 1962-1963. Because his career during this time took him across the country, I was interested to see what types of subjects caught his eye in each region. To do this, I placed each picture into a category based off of the provided subject descriptions in the data set, which were taken from Cushman’s notebooks. Each point is clickable, and upon doing so a window pops up with some of the descriptive data available in the full data sheet, as well as a link to view the picture itself. Many points are overlapping, and can be cycled through by clicking the left or right arrows in the upper corner of said window.
In terms of concrete value, this method of mapping and categorizing data could be useful in gauging what it is about certain locations that people value, or tracking the interests of one person over time. For instance, the photos taken in the Southwest are primarily of landscape features, which could suggest that Cushman found these features to be exceptionally interesting or novel compared to those in other regions he visited. Meanwhile, the southern Midwest has many blue points, representing buildings and man-made structures, which could indicate that the architecture in this region is noteworthy in some way. Especially in the age where almost every picture we take is geotagged, this data could be used on a very broad scale by park managers (which areas are people most interested in?), tourism services (how are people utilizing aspects of a space?), state advertising boards (what aspects of a state do non-locals find most exciting?), and more.
The data I used to create this map came from the Indiana University’s Charles W. Cushman Kodachrome Slides collection, a digital archive of the works of photographer Charles W. Cushman. The coordinate data was drawn from Google Maps and displayed using ArcGIS Online.
The process of cleaning the data was rather grueling for three reasons: first, the original csv file contained nearly 70 columns of information for each of the 340+ data points; second, many of these columns were empty, which made finding a consistent method for locating each point difficult; third, there was no coordinate data that could be used for mapping, since the data was stripped directly from the Cushman’s notes taken in the 60’s. The locations for each point ranged from addresses to highway names to entire states (or none at all), which made finding coordinates extremely time consuming. After finding coordinates for each photograph, I noticed that there were far too many categories to map effectively, so I read through the descriptions of each picture and placed them into 11 general groups: plants, weather, landscapes (mountains, forests, deserts, etc), people, buildings, boats, public spaces, animals, aquatic landscapes (rivers, shorelines, etc), aquatic infrastructure, and fireworks (there was only one fireworks picture, which fit into no other category).
In terms of web design, I wanted to make my project as simple as possible. This is because of my own past experiences with poorly designed projects and a general lack of design savvy.
One significant pitfall of this method is that, because of the lack of location specificity, many of the points overlap by 100%. This obscures the visual clarity I had hoped the mapping method would provide, and ultimately makes discerning a pattern in high-density areas impossible. For example, one point in San Francisco encompasses over 60 pictures while only appearing as a “plant” category. While there are tools to solve this problem using the desktop version of ArcGIS Pro, Carleton does not pay for student access and I was therefore unable to utilize them. Having to create my own coordinates for each point was also extremely time consuming, and prevented me from dedicating more energy to the visual presentation of this project. In conclusion, things would have been much smoother had the data been images from someone’s iPhone.
