Machine learning to predict neighbourhood change

A new machine learning prototype developed by UNSW researchers can help predict neighbourhood change and the subsequent socio-economic impacts on a region’s population.

The model has been developed and tested by researchers at UNSW City Futures Research Centre and aims to better equip policymakers and government with the knowledge and data required to deliver more equitable city planning and outcomes. Initial findings were published in the Cities journal.

The paper focuses on gentrification — an urban phenomenon affecting many cities around the world and recognised as a type of neighbourhood change that sees poor or historically working-class neighbourhoods experience drastic changes to the demographic, land use and housing affordability.

Gentrification can displace and negatively impact residents such as low-skilled workers and vulnerable people, and governments and policymakers often struggle to address the associated harms.

“Gentrification is often identified when it’s too late, and it can be costly to address the harms it has caused,” said William Thackway, postgraduate researcher at UNSW City Futures Research Centre.

“The ability of policymakers to adequately tackle harms caused by gentrification rests on proactive strategies which prevent or mitigate displacement of vulnerable people before it becomes too expensive to do so.”

Gentrification in Sydney

Thackway, Professor Christopher Petitt, Dr Matthew Ng and Associate Professor Chyi Lin Lee developed the prototype machine learning model and tested various data from Sydney as a case study.

“A key finding from our work is that the gentrification frontier is predicted to expand outwards even further from the city centre,” Thackway said.

“Previously, the rings of gentrification were in 5- to 10-kilometre rings around the Sydney CBD, but that is predicted to expand to 10–20 kilometres.”

The study identified an effect the researchers call “spill over” as an across-the-board indicator of predicted gentrification in Sydney. Spill over is when displaced residents from gentrification hotspots move to neighbouring suburbs where rents are slightly cheaper.

“In the early 2000s–2010s, the Inner West became a major hotspot for gentrification, but our analysis predicts deeper Western suburbs like Auburn and Bankstown could be new hotspots,” Thackway said.

Eastwood beyond Ryde and Brookvale beyond Manly were other suburbs experiencing spill over effects and predicted to gentrify according to this machine learning analysis. The analysis also reaffirmed Botany and Maroubra beyond Randwick as gentrification hotspots.

Besides spill over effects, other potential drivers of gentrification and neighbourhood change in Sydney included changes in crime levels, new transport developments and further changes in demographics.

Image: Map of predicted neighbourhood change in Sydney between 2016 and 2021. Areas in orange/red predicted to gentrify, areas in purple/blue predicted to decline in socio-economic rank. Image supplied by Thackway et al. (2023).

Surprising indicators of gentrification

The advantage of this new machine learning model is that it can make links between variables that are otherwise overlooked in other methods of analysis involving just human expertise.

“Our study includes a wider range of predictor variables than previous machine learning studies, spanning socio-economic, housing, business and Airbnb data,” Thackway said.

The machine learning model was trained and tuned using over 80 predictor variables from a wide range of data inputs such as property reports, the census, business registry and Airbnb.

To test its accuracy, the researchers retroactively applied the model to previously ungentrified neighbourhoods that ended up becoming gentrified.

“Family compositions and relationship status were surprisingly important indicators of gentrification in some areas of Sydney,” Thackway said.

“It was surprising to see that an increase in married couples in an area led to a higher prediction that the area will gentrify, while areas with more divorcees and one-parent families were less likely to gentrify, according to our model.”

In some cases, family and relationships were as important as house prices, education and employment in predicting gentrification for a suburb.

Towards better quantitative methods

Predictive modelling and machine learning tools in the urban policy spheres are still in their infancy.

“There is still scepticism among policymakers about the trustworthiness of such models,” Thackway said.

“Previous machine learning models have had a ‘black box’ element to them, meaning that we can’t see how machine learning came to its conclusions. Because of this, the preference among policymakers is dominated by qualitative methods.”

But this new machine learning model developed by UNSW researchers can predict gentrification with 87.3% accuracy and it eliminates the ‘black box’ element by implementing a model explanation tool that interprets how the machine learning model came to its conclusions.

“Qualitative methods like the Neighbourhood Change Warning System and Gentrification Index are easy for policymakers to understand,” Thackway said.

“But the downside is that they are quite simple and lack robustness. Our machine learning model incorporates tens, if not hundreds of indicators compared to qualitative methods. The advantage of using machine learning as opposed to basic indicators in qualitative methods is the model can identify interactions and relationships between variables that one might not necessarily be able to do just from human expertise.”

Overall, the UNSW team created a more holistic, robust and explanatory machine learning model that improves best practice for predicting future gentrification hotspots.

Opportunity for future use

The tool is in its development phase and there is scope to test it to more extreme degrees to ensure its performance.

“Right now, the major implication of our work is that this model can produce meaningful and powerful results that will enable proactive policy decisions and interventions for urban planners,” Thackway said.

“While we used Sydney as a case study to test the model, it can be applied to comparable cities by inputting similar data.

“In the past, most gentrification research has looked at what’s already happened to analyse drivers of gentrification. This machine learning model enables predictive modelling of gentrification.

“With the 2021 census data release upcoming, forecasting vulnerable areas for 2026 will provide policymakers with an empirical tool to proactively intervene and design more equitable solutions for vulnerable communities.”

Image credit: iStock.com/MF3d