Data in Urban Planning: Data Science Internship at Digital Planning Lab
Have you ever wondered how land-scarce Singapore is being planned? Are you passionate about using data to solve real-world problems?
As a Geography major, I am interested in the social, political and environmental aspects of urban spaces. However, in today’s world, digital technology is integrated into every industry. Learning to code is almost as important as learning to read and write. Therefore, at University and in my free time, I have spent some time learning to use Python and R to analyse data and automate simple tasks.
Over the 2020 winter break, I was able to pursue my dual passion in urban planning and data analytics. For 5 weeks, I was a Data Science intern at Digital Planning Lab, a technology-oriented department within Singapore’s Urban Redevelopment Authority (URA).
What did I do during the internship?
During my internship, I was tasked with analysing Singapore’s public transport transaction data to gain insights on the transfer routes (e.g. changing from Bus to MRT to get to a destination) in Singapore. Coming from a non-technical background, I initially struggled to efficiently process the gigantic dataset. However, with some tips from my supervisor and some googling (every programmer’s best friend), I was able to properly clean and transform the dataset using Python’s Pandas library.
Have you ever taken a train to an MRT station and then transferred to a bus in order to get to a destination? One of the biggest challenges I faced during my internship was to improve the current methodology for classifying if a trip should be considered as a transfer (the example above). I had to examine the dataset and experiment with various statistical methods (thankful for my statistics modules) to calculate a cut off duration.
To further refine the classification method, I had to use my personal experiences and prior knowledge of Singapore’s public transport. For example, if a commuter is alighting and boarding from Orchard MRT within 10 minutes, they more likely to be engaging in a short activity (e.g. carousel transaction) than transferring. Hence, I had to incorporate an understanding of real-life scenarios and insights derived from big data in order to develop a comprehensive, hybrid methodology.
After processing the dataset, I identified the most used transfer routes in Singapore and created a map displaying the hotspots.
From the map, you can identify the regions and towns which have a higher density of transfer trips. Unsurprisingly, since Singapore’s public transport system was designed with a hub and spoke structure, most of the hotspots are near prominent MRT stations. Notably, this method excludes transfer trips within MRT stations (no exit required) such as transferring from the North-South Line to the Circle line within Bishan MRT.
Thereafter, I conducted a network analysis on individual towns using the walking network dataset. This allowed me to identify the various pathways that are most likely to be used by commuters who are transferring. While commuters who are transferring are not the only users of public pathways, previous ground surveys have shown that they are one of the largest group of users. These pathways could then be recommended for future improvements to enhance pedestrian comfort such as building shelters or widening pathways.
Other than my project, I had the opportunity to better understand how Singapore is planned by attending an engagement session with Singapore’s Chief Planner. Furthermore, I was invited to URA’s GIS Day, where I learned about the various digital tools used by Urban planners.
What did I learn from the internship?
This internship made me realise the complexity and difficulty of analysing data for real-world insights. In school, datasets are generally cleaned and directly suited for generating insights. However, in the real world, datasets may not be directly suitable and may require a more robust methodology and further manipulation to yield relevant insights.
The internship has significantly improved my competencies and confidence in analysing data using Python. Furthermore, I realised the importance of writing efficient algorithms. Previously, when I was working with smaller datasets, I would usually use the most straightforward approach and ignore efficiency. However, with a large dataset, the straightforward approach might be too slow and might even cause your computer to crash.
Most importantly, I have gained a deeper understanding of how big data is gradually informing and enhancing urban planning and public policies in Singapore.
Concluding Notes
Despite the internship only lasting 5 weeks, I had thoroughly enjoyed my time at Digital Planning Lab.
I am really grateful for my supervisors, Songyu and Zhongwen for being extremely approachable and guiding me throughout the journey. I am also thankful for my fellow interns, Denyse, Jia Xin, Lawrence and Jim for helping me and making the internship an enjoyable journey.
Feel free to contact me on my LinkedIn if you would like to find out more about my experiences at the Digital Planning Lab.
References: