Clock Icon - Technology Webflow Template
15
min read

Assessing Liveability through Walkable Neighbourhoods

Governments can create vibrant communities by using our 'liveability' scoring system for suburbs

20-minute neighbourhoods are a policy and planning initiative supported by the Victorian government under PlanMelbourne 2017-2050. The concept promotes ‘living locally’ by encouraging infrastructure design to equip residents with the ability to access most of their daily needs within a 20-minute walk from their home.

DiagramDescription automatically generated
Figure 1: Source - https://www.planning.vic.gov.au/policy-and-strategy/planning-for-melbourne/plan-melbourne/20-minute-neighbourhoods

Empowering residents by providing them with local social infrastructure doesn’t just create vibrant local communities it also reduces traffic congestion and pollution while facilitating community health outcomes by encouraging active transport (walking and cycling).

Because 20-minute neighbourhoods have such a tangible impact on the liveability of local communities, we at propella.ai have designed a liveability framework to evaluate locations by understanding their access to basic social infrastructure within a 20-minute walk.

Acquiring the data to measure liveability

The first step to assessing the liveability of locations through walkability is data acquisition. This involves defining a 20-minute walking catchment around a site, by tracing the shape that a resident can walk along the footpath network in any direction from the site in up to 20-minutes.

MapDescription automatically generated
Figure 2: 20-minute walking catchment around an arbitrary location (turquoise line) and the theoretical uninhibited walking area, 1600m radius (grey dashed line).

Using the site’s walkable catchment, core amenity types are independently assessed with two measurements:

1.     Infrastructure abundance - The number of accessible amenity sites within the 20-minute walking catchment;

2.    Proximity -Minimum walking time to the nearest amenity site along the footpath network. 

The core amenity types assessed in propella's liveability model include:

* Walkability is measured as the ratio of the area that can be walked along the footpath network, divided by the theoretical area around a point (1600m radius) if there were no impediments(e.g. roads or geographic obstacles). This ratio is commonly referred to as a ‘ped shed’.

** Parkland is assessed using total sqm of parkland within the 20-minute walking catchment and proximity to nearest park.

Figure 3: Screenshot from propella's location intelligence app, showing public transport stops for a given 20-minute neighbourhood, including walking time to nearest train station

 

Comparative Analysis of Amenity Data to help Assess Liveability

Once the data for core amenity types has been acquired for the site, it is compared against a pre-processed sample of over 50,000 addresses around Greater Melbourne. We use this representative sample to empirically rank the addresses in the sample across every data point.

The empirical ranking is a key focus in our liveability model because it allows the scores to be anchored by what physically exists in the built environment. This methodology has the added bonus of scores that express which quantile the address belongs as part of the score. For example, a score of 0.81 indicates that 19% of the sample scores more preferably for that measurement and 81% has a less preferable measurement.

Another feature of empirical ranking is that over time, as changes in social infrastructure are recorded within the amenity data (either additions or removal of amenity), liveability scores will be re-calculated for the sample of walkable neighbourhoods.

Figure 4: Distribution of accessible parkland in the 50,000+ address sample

Figure 5: Distribution of index score after empirical ranking

In the Figure 4 above you can see the distribution of accessible parkland (sqm) within a 20-minute walk for the sampled addresses. It can be observed from the distributions that most addresses have access to less than 500,000 (sqm). Very few addresses have access to a lot of parkland as exhibited by the distribution’s long tail - the tail indicates what is referred to as skewed data.   

When empirical ranking is applied to the distribution (Figure 5 above), you can see that the “skew” in the data is removed. How much more accessible parkland exists from one address to another becomes irrelevant. The only distinction is how many locations in the sample have more parkland. By performing this ranking on the data, subjectivity about how much parkland is required to make a location liveable is ignored.Instead, the unit of measurement is how well serviced the location is with parkland when compared with the rest of Melbourne, which is not a subjective measurement.

Deriving a Liveability Score

Our liveability model attempts to remove as much subjectivity as possible from the analysis. However, subjective weightings need to be applied to each measurement data point in order to combine them into a single overall score. Measurement scores are aggregated into 6 components of liveability:

The colours of core amenity types assessed, and liveability components, link which measurements that were used to assess each component. The sample is empirically ranked for each component.

Chart, sunburst chartDescription automatically generated
Figure 6: Weighting contribution of each liveability category to overall liveability scores. Contributions total to 100%.

Figure 6 outlines the contribution of each component to the overall liveability scores. Parkland and access to public transport have the largest contributions (25%). Despite walkability having the smallest contribution (5%), the walkability of an area has consequence for all other categories. This is because the site catchment is defined by a 20-minute walkable area, if the walkable area becomes larger the site catchment has a greater area to capture additional amenity for each category.

The final output is a dataset with that empirically ranks locations around Greater Melbourne using 20-minute neighbourhoods as the focal point of analysis.

Figure 7: Liveability scores around Greater Melbourne. A score of 1 is assigned to the most liveable location, while corresponds to the least liveable location

Using the Liveability Indices

Our liveability assessment methodology provides a quantitative means of measuring how liveable a given site is, within the context of the 20-Minute Neighbourhood (20MN) policy framework and the greater city the site is located in. But what is a practical application of the Liveability scores?

Developers of large residential or mixed-use developments can undertake a baseline analysis of the 20-minute neighbourhood for the development site.  Then the services and amenity that are being introduced as part of the development (related to the social infrastructure criteria within our liveability model) can be added to re-calculate the liveability indices including the new amenity.

Figure 8: Liveability scorecard for a proposed residential development after addition of new social amenity

Figure 8 shows the Liveability indices for the 20-minute neighbourhood for a proposed development site.  Accounting for the addition/release of substantial parkland on the subject site, improvements to walkability, and additional of new retail shopping services on the site, the overall Liveability score of the 20MN for the site improved from 7.6 (out of 10) to 9.2.  This puts this site in the top 8% of 20MNs for Greater Melbourne, an attractive proposition for future residents!

We have partnered with CoreLogic's "Onthehouse" platform, to determine Melbourne's most liveable suburbs! You can read about that project here @ propella.ai/blog/new-liveability-rating-score-designed-to-provide-consumers-with-insights-where-to-buy-and-invest

You can read more about how we devise "liveability" scores using "walkability" in our blog @ propella.ai/blog/measuring-the-20-minute-neighbourhood

For further information on this subject, please contact us @ https://propella.ai/contact

Matt Molony

Head of Data Science

With a background in Applied Mathematics and Engineering, Matt leads data science at propella.ai, developing geospatial analytics and deploying AI and machine learning models. He also manages the technical team, guiding them on new methodologies and tools to drive business success.