SPR Webinar: New Data, New Insights: How new data sources and techniques are changing property research
18 July 2022
Potential and pitfalls
This webinar explored a number of angles on a topic that moderator David Inskip of CBRE Investment Management admitted is a very broad one, encompassing as it does areas like big data and open data. But the importance for property researchers in their day jobs will be in how it can help their organisations to make better decisions.
Dr. Chlump Chatkupt of PLACEMAKE.IO touched on some of this potential with examples of novel analytical techniques that can harness existing data in new ways or access new sources of data. One possibility is analysing rental growth based on railway station catchment areas, which intriguingly can sometimes mean higher readings nearer to the station, but sometimes lower. In another example of highly granular locational analysis, he showed ways to identify dynamic data variables that could identify micro areas of London that are ripe for gentrification. And a data source that few would have considered for real estate prospects is government-generated indices of deprivation - these can give an indication of where there is potential for commercial improvement in a locality.
In his short presentation, Simon Hayter of Knight Frank then highlighted some aspects of property-related ESG data that are evolving rapidly, particularly in the ‘S’ and ‘G’ space, where trends are often seen as less quantifiable than for environmental factors. He noted that while some of the social data relating to buildings is demographic or perhaps transport-related, collecting information on individuals’ behaviour can soon run into privacy issues.
This was a key theme of Tom Wainwright’s presentation, too. The Associate Professor at Royal Holloway sees this not just a potential problem for members of the public but also for commercial organisations in terms of reputational risk. He observed that in the ongoing data ‘gold rush’, which has a primary goal of making better financial decisions, ethical dilemmas can get overlooked. While the data trails that individuals leave may potentially provide some highly useful analysis, for instance on spending habits in different locations, the individuals concerned are often unaware of how this data could be used. AI models can also sometimes build in discriminatory patterns, which mean that the conclusions they generate may not be objective.
In the panel discussion that followed, the question of who owns the data created by individuals inside buildings they are visiting emerged as an ongoing bone of contention, with no clear consensus on whether it is the building owner, the individual or the organisation ultimately holding it. Responsibility for the accuracy of data is another minefield, as there are no consistent standards in place across the board. Wainwright suggested that a lot could be learnt from the open data community, where there is a constant questioning of the rights and standards to apply, while Chatkupt emphasised that data providers should rigorously test the statistical foundations of their models, with the relevant technical discipline brought to bear.
There was however agreement that using the most appropriate data depends on asking the right questions in the first place and not just generating masses of data for its own sake, which has often happened with proptech start-ups. This is clearly part of the role of property researchers, whose place should not be seen as under threat from a deluge of AI driven data, although there may well be a need for greater understanding of data science. Likewise, the ‘expert professionals’ who tend to dominate the real estate business will continue to be important. Hayter argued that data cannot always replicate experience in such a complex field.
Tim Horsey