Ideas and Thoughts
Slopes, a hidden weight of Accessibility
Historical Context of NUS' Main Campus
Source: Twitter @sehof
I stumbled across an interesting tweet - one that resonated with me as a student in NUS.
To summarize the post, student activism was a heavy concern back in those days and the architects designed the NUS' main campus to suppress "disruptive possibilities". The plans seemed to revolve around NUS' rather slopy topography. Amenities, which were central to student life (e.g. student centres, NUSSU, hostels), were strategically placed in disadvantageous locations to prevent students from congregating.
Throughout my 3 years studying at NUS, my friends and I have always felt this to be true. Everywhere on Campus that's worth travelling to was in fact a huge pain too.
Important pathways seemed to require us to scale a massive hill; the most "central/accessible" locations tends to have the highest slope angles. The Main Campus is extremely pedestrian-unfriendly and the internal shuttle seems to be the only sensible mode of transport. I find it insane that the NUS central library (CLB) is located on the highest point of the campus and it is unavoidable to climb numerous flights of stairs/ steep slopes in order to reach the building no matter which faculty you start from. This tweet reassured me that there was some sort of logic to what I experienced.
Comparing UTown with the Main Campus
Source: My Slope Model (DEM) generated by Google Earth
Source: STX Landscape Architects
I have always noted the pedestrian-unfriendlienss of the main campus mainly because I was a resident of UTown and I often try to reason why the experience was starkingly different even though the topographic characteristics were similar. The above figures show the unevenness of the plot of land, visualizing distinct ridges within both plots of land.
Now that I think about it. UTown's landscape design is pretty impressive, isn't it? I've been living in RC4 (a student hostel) for 2 years. I've enjoyed taking walks and exploring different nooks and crannies of the compound (I miss the benches by the Berlin Wall site). There is also a multitude of hangout spots were left a lasting impression on me - some nights I sit by the grass slopes and stairs to pet the resident cats. It's pretty inspiring to me how the landscape designers have integrated the pathways and open spaces into the texture of the place.
Source: Photo © John Gollings Photography Pty Ltd
Project Details of the Architecture can be viewed here.
Beyond being pedestrian-friendly or pedestrian-unfriendly, the difference in the pedestrian experience between UTown and the Main Campus reflects the distinct functionality of each space based on a their respective historical context. More interestingly, this difference was accentuated through strategic planning of slopes and accessibility - two simple metrics that I felt was well manipulated to fit its purpose. Indeed it is pretty cool to see such a juxtaposition.
Research Question:
Can this narrative be captured easily by current geospatial/topographic data?
- A exploration of accessibility and slopes angles within NUS
This further makes me question - can this difference in spatial design be captured quantitatively as data?
With a growing demand for evidence-based design research, I felt this was an interesting preliminary study to explore more physical metrics that could influence people movement in a spatial/urban environment.
Space Syntax (a community I align to) studies physical spatial characteristics like angular change as they theorize that pedestrians tend to walk in straight lines and influences their decisions in walking certain paths. This has been integrated into the fundamental logic in their network models to forecast movement patterns. In my humble opinion, I always thought that slope angles (vertical angular change) of paths are equally important. As inferred from the narrative I've described, people should and are discouraged by slopes.
So why are people not studying it even thought it seems seemingly obvious? Well, slopes have been something that have been difficult to study since mapping technology from past to present have been largely 2D and there is a lack of data. Also, it is uncommon for most urban sites to have hilly topography in which the slope angles of pathways might be influential in the prediction models. Hence, this conversation has been avoided because of such practical limitations. However, I thought it'll be nice to have this headstart, mainly because I'm curious and I thought UTown/Main Campus will make a very appropriate case study.
Broadly speaking, I hypothesize that the more "accessible" pathways in NUS main campus tend to be be slopy to discourage students travel to central areas to congregate. On the other hand, the converse might be true in UTown where "accessible pathways" are gentler and slopes are avoided in order to ease movement within the residential community space. Hence, by studying the relationship between slope angles and accessibility of pathways in both sites, I hope to illustrate a more complex account of the respective pedestrian experience.
Methodology
%20GPS%20Data%20from%20Google%20Earth_JPG.jpg)
%20DEM%20and%20Contour%20Representation_JPG.jpg)
2) Converting them into a Digital Elevation Model (DEM) and visualizing them as contours
%20Slope%20Analysis_JPG.jpg)
3) Slope Analysis; Approximating data as angular change as opposed to elevation
%20Street%20Slopes_JPG.jpg)
4) Pushing these data onto the street network model (red is slopier/black is flatter)
5) Our street network model has already computed metrics of accessibility - choice and integration on multiple scales (400/800/1200). So we shall explore correlations between these metrics and the slope angles!
Results
*Dotted Line represents the median value
Through-Movement (Pathway as a Thoroughfare)
To-Movement ("Center-ness" of Location)
Radius of Movement
400m 800m 1200m
Short Walk Medium Walk Long Walk
Discussion
To understand the relationship between slope and accessibility, I drew dotted lines (representing median values) for both quantities to compartmentalize them into a matrix. Its purpose is not so much to "count" the number of pathways in each section, but rather to observe the unique distribution.
In the main campus, it is observed that the slope/accessibility of pathways within seems to be very evenly/normally distributed. This can be inferred by how centred the median lines are in the graph where the points spread radially from the midpoint. There seems to be an equal chance that the slopy pathways are low and high in accessibility - it's as if the campus was planned without any real regard for the topography.
The "UTown" Effect
UTown, on the other hand, appears rather interesting. Right off the bat, we can already see a disproportionate distribution of slope/accessibility in its pathways. This is characterized by the huge area above the median line for slopes (1&2 on the matrix) which shows that pathways in UTown seem to be on two extremes, either very flat or very slopy.
Taking a closer look at the most "slopiest" paths in UTown, they appear to fall into the dotted purple box, where it is typically located at the top left-hand corner of the graph. This shows an interesting pattern - the slopiest pathways in UTown are ALSO the least accessible, be it in the form of "Through-Movement" and "To-Movement". It seems that this was specifically what the architecture team of the UTown project meant when they advocated for "space and place-making while respecting the natural contours of the site" in the public profile.
Ending Notes
I'm rather satisfied with this short self-initiated project conducted over the school semester. These findings, even though a little unprocessed and raw, support my initial hypothesis where the relationship between slope angles and accessibility of pathways explains the difference in the pedestrian experience. I think more detailed studies can be done to explore this concept further like to understand how slopes affect people movement on varying spatial scale (I would think stairs matter a lot more if your journey is longer). However, due to the low reliability of my data, I dare not use it in such means.
In a way, this still further asserts the possible need of capturing slope data in the future as I believe it unveils a hidden dimension to pedestrian forecasting that we have often tended to overlook. Also, please tell me if anyone's willing to volunteer walking the NUS campus to collect more detailed data of the space.