How We Use Data To Ensure Urban Planning Meets Community Needs

Optimal Urban Planning Requires Big Data and Deep Analysis

‍

Parks and open green spaces are sometimes an underrated part of our communities, when it comes to urban planning and how we live our lives. Where places such as supermarkets, doctors and train stations are important for making our communities function, they can also be very one dimensional in the fact that they only provide one core use. This is not the case for parks.

‍

Caption - Albert Park Lake, Albert Park, Melbourne

Using People Movement Data from Parks to Support Better Urban Planning and Development

As a community we all use parks in many different ways, ranging from taking the kids to kick the footy to a change of scenery while studying/working or perhaps even catching up with a secret date that you would prefer to keep on the down-low. Particularly during the times of the COVID lockdowns, parks allowed many people to leave their houses, clear their heads or see a friend that would have been impossible otherwise.

In this article, we take a look into a few ways we can analyse activity in parklands, and how this information is useful for local Government and Urban Planners to help improve community amenity and social infrastructure.

The data used for this analysis is Mobile Device Location Data (mobile data for short). This data is collected through apps on our phones and shows where a device is located at a particular time. While people can get spooked by privacy concerns when considering this concept, mobile data is highly anonymised and never used to follow a single device’s movements, more of which can be read about here https://www.propella.ai/privacy-policy

The data in its raw form is close to unusable to us, so our data scientists have developed a pre-processing of the data to break it into what we define as ‘Stays’ and ‘Travels’. This provides highly valuable data-led insights for urban planning or development.

Stays are defined as a geographical point at which a unique device has been determined as stationary for a defined period of time or longer. A single stay event is made up of multiple raw data events where the device stayed relatively sedentary for the duration of the stay. Stays data is used to find places that attract users for a specific purpose.

Travels, on the other hand, are identified as a geographical point where a device was identified as being in motion. A single device can have many travel events across different time periods in different locations. These events help us see the paths users take to traverse a particular area.

Both the ‘stays’ and  ‘travels’ are used in different ways in order to analyse the usage of parks and can also be used together to draw insights.

Using Stays Analysis Data for Urban Planning

Due to the sedentary and more intentional nature of the stays data, our analysis is conducted firstly by looking at the natural groupings, or ‘clusters’, of the given events. At the foundation of these clusters is the fact that people tend to gravitate towards different areas in the parks (as we see in the image below). This could be for many reasons such as the amenities (BBQs, toilets, tables etc.), large open spaces (good for active activities such as frisbee) or even just for the beauty of the area. These uses are the next part of analysis we investigate.

While the actual activities occurring in the particular clusters is impossible to deduce from the stays data alone, other statistics can then be drawn from the data, allowing a keen analytical mind to draw inferences regarding the area. From the clusters, statistics such as time of day usage, day of week usage, average duration of stay and more can all be analysed to help infer the usage of a certain area.

For example, a common pattern seen around football grounds and other sports facilities is a spike in activity on Saturdays (game day), with another spike in activity at some point during the week, generally Wednesday or Thursday evenings (training night). Cafes also show very characteristic distributions of heavy lunchtime traffic, with reasonable numbers throughout the week but spiking on the weekends.

Below we see a couple of maps of parkland around Albert Park Lake. On the left, we see the raw data from mobile data events, while on the right we can identify clusters of activity. This map indicates people "stay" and spend time around: Melbourne Sport and Aquatic Centre (cluster in lilac colour); Melbourne Sport Centre Lakeside Stadium (cluster in purple colour), Albert Park Indoor Sport Centre (cluster in red colour) and more.

‍

Caption - Figure 1: Albert Park Lake parkland maps. The left image shows the raw data from mobile data events. The right image reveals clusters of activity

Using Travels Analysis Data For Urban Planning

When analysing the travel events through a park, we are often looking for where people are moving to get from one point to another and what the major ‘ant trails’ are that are formed by these paths. In this application, it's far better to look at the parks much more as a whole network to see how people move from point to point through the park.

Using the Albert Park example once again, we can see (below) instances of travel events occurring within the park. While some walking activity may occur along the roads, due to the size of the roads in Albert Park they have been marked in black to distinguish between probable pedestrian/cycling traffic and likely vehicle traffic.  

Many ‘ant trails’ can be seen in this map most notably the trail around the perimeter of Albert Park Lake and the smaller roads connecting many of the sports grounds in the southern-most section of the park. This gives us very good indications of how people are moving through the park and these different sections can be further analysed if desired.

‍

Caption - Figure 2: Instances of people "travelling" within and around Albert Park Lake

Why is this Data Analysis Important for Urban Planning?

There are so many uses for this type of data in the analysis of parkland and open spaces. In its most basic form, as we have shown through this article, we can draw many inferences as to the usage of different areas of the parks and, armed with this information, councils can make informed decisions on how to improve areas in a way that is better tailored to its existing usage.

Other insights lean more towards identifying an area of the park that may have shown up as a cluster during the stays analysis, however, appears far more disconnected in the travels analysis. Actions such as investing in a path may be taken to create a more connected and flowing space.

Insights vary from park to park and city to city, but this data has been powerful for many of our clients to date. We have helped them assess changes in activity levels before, during and after COVID lockdowns, to help evaluate which parks and open spaces require greater capital investment to support community needs due to the impact of COVID.

If you’d like to learn more about how our location intelligence platform can assist you to better understand how public spaces are utilised, please get in touch with our location intelligence team at www.propella.ai

‍