SIGNAGE CONDITION
The University of Utah has compiled a complete set of data about all transportation-related signage on campus. As part of that set comes a one-word description of each sign’s condition – good, fair, or poor. This study looks at a subset of these signs by analyzing the western area of the University’s “Main Campus” to see if the age of nearby buildings has a positive correlation with the condition of the signage.
GEOG 6000 - Advanced Geographical Data Analysis
Project Background
In the summer of 2019, the University of Utah conducted its first campus-wide pavement marking and signage inventory. The primary objective of the inventory was to figure out the exact number of signs and markings that are owned by the University and create a maintenance plan to keep them in good working order.
After the inventory itself was completed, the data became available for studying. This study began as an attempt to look at the conditions of signage across campus and find a possible correlation between that and some other factor. The signs were originally designated to be in good, fair, or poor condition, and this looks for a potential reason behind why some areas have signs in better shape and other areas do not. The factor that became the focus of the study was the age of nearby buildings. As the college has been running for over a century, there is quite a wide range of building ages on campus. The hypothesis became that signs where the nearest building was substantially older would be in much worse condition than signs where the nearest building was relatively new. As such the statistical analysis comparing the conditions of signs with the nearest building’s age began.
Study Area
As has been established, the study area is the western part of the University of Utah campus. This area was selected as it provided a very wide range of buildings ages with very dense signage. The study area contains the oldest part of campus, President’s Circle, as well as some more middle-aged buildings and one of the newest buildings on campus, the Law Building. Signage is abundant in the parking lots and in the streets, but there is also some in the interior of campus too.
Figure 1: Study Area
Sign Locations
All three condition locations can be plotted as separate figures, seen here. This indicates that the spatial density of the signs in good condition are the highest amongst the three, while the poor signs have the lowest density.
Figure 2: Sign Locations
Kernel Density Heatmaps of Each Signage Condition
Figure 3: Kernel Density Heatmaps of Each Signage Condition
Relative Risk Models for Each Condition
Relative risk models were created, which is the ratio of the probability of one event occurring in one group versus a different event occurring in another group. In this case, it is the likelihood of a sign in a certain condition appearing in a certain location.
Figure 4: Relative Risk Models for Each Signage Condition
Campus Building Locations & Age
The next step in the analysis is associated with the the age of the buildings on the campus. Finding the ages of campus buildings was the most difficult part of this process as it involved looking through historical records. Once they were acquired, a point feature class was created containing the centroid of each building as well as the age. From there, the feature class simply needed to be re-projected. The buildings were then plotted based on their location and age, seen here.
Figure 5: Campus Building Locations & Age
Final Relative Risk Model
The last stage of the analysis is to combine the signage and the buildings into something that can reveal whether or not there is a true correlation between older buildings and campus signage in poor condition. The first step is to build a model that shows the variation between the points using the covariate, which in this case is the age of the buildings. This model takes into account the signs, the building location, and the building ages. Then another model is created that does not use the covariate but only involves the signs and the building locations. Running an ANOVA test determines that the first model with the covariate is the more fitting model, as the p-value is quite large. Finally, since the first model fits better, one more relative risk analysis is run to show correlation between signage conditions and building ages.
Figure 6: Final Relative Risk Models
Discussion & Conclusion
Signage at the University of Utah is largely in good condition. This subset of the larger study overall proves this fact 188 signs in good condition, 53 signs in fair condition, and 48 signs in poor condition. However, it is important to consider the importance that signs have on our campus, especially for new or prospective students. They can be a reflection of the campus at large, and may have an effect on if certain students decide to come to go to school there.
This study asked the question if certain areas are more prone to have signage in poor condition - nearby building age was presented as a possible answer. Building age obviously isn’t the only factor that would affect whether or not a sign is in poor condition, but it says a lot about how long it has been since certain areas were under development and signage was being installed. The kernel density models do show that in many cases signs in each respective condition tend to be clumped together, particularly those in good condition. However, signs in fair and poor condition may be slightly more spread out. Building ages show a strong clustering as well, as all of the oldest ones reside in the historical part of campus. The last figure that takes everything into account clearly shows that the area of campus with the oldest buildings is far more likely than the rest of campus to have signs in poor or fair condition. Building age appears to be a positive correlation in this study area, and it would be interesting to see the results of such an analysis for the campus at large, given more time for research.
Building and developing signs on any campus is always going to be a fluid process. Designs will change, needs will vary, and conditions will slowly get worse. But narrowing why the latter of these factors is occurring is critical to understanding when and where signage needs an update. This study visibly displays nearby building age is an indicator for whether or not a sign is likely to be in good or bad condition, a factor that will be important to for decision-makers to consider when creating the campus maintenance plan mentioned in the introduction. This knowledge will help alleviate the need to perform one of these inventories on a regular basis and potentially improve the way campus funding is spent.
Skills Used / Learned
GIS Workflow
Spatial Analysis
Project Design
Model Building