Tenure Intelligence (Ti) - Combining artificial intelligence and housing data

Multinational technology companies are not the only ones using big data and computing power to make predictions about the unknown. Councils have seen the benefits of machine learning and are now starting to adopt this approach to make public services more effective and productive.

I’ve been fascinated by the data councils hold on residential properties and how it might be utilised for many years (I know, very sad). An opportunity to put this interest to the test came in 2012 when I was challenged with identifying 10,000 unlicensed and unsafe private rented properties during the set-up of Newham’s ground-breaking borough wide licensing scheme.

The traditional solution to this type of problem was to dispatch a small army of council officers to walk the streets, spotting properties that ‘looked’ like they might be rented, through signs of disrepair, overflowing bins and even the ‘dirty curtain test’. On a sunny day this was quite nice task for council officers, however it’s a costly and unreliable approach, particular across whole boroughs . This raised the question, could I find indicators of property tenure in council data and use it to make predictions about tenure?

Jumping many steps/years forward, with the support of some clever colleagues and dedication to refining the technique; Ti was born. We have now shown that big data and advanced mathematical models (also known as AI) can be deployed across councils with strong results. Offering a much cheaper and more effective solutions which turns out to have many beneficial side effects.

How does Ti work?

Ti uses pools of council-held data to identify trends at the property level. Mathematical models are trained to recognise data trends for each tenure type in each council area. These data patterns are then used to predict tenure and other factors for all residential properties.

The key stages of Ti are:

• A wide range of council-held data is assembled on council systems using unique property reference numbers (UPRNs) as data keys
• Data is analysed by skilled and experienced housing practitioners using bespoke and locally built algorithms to produce risk scores for each residential property
• Every effort is made to remove error from the mathematical models to achieve the most accurate outcome possible for each council
• Risks scores for each property can then be analysed at a macro and micro level. This includes a final report where required for public consultations and Cabinet reports etc
• Councils are left with the ability to isolate groups of properties of interest on a street by street basis, including HMOs or properties with Category 1 hazards etc.

What is the value of Ti?

Ti provides local authorities with a new tool to help differentiate between properties which are privately rented, HMO and those that are owner occupied or socially rented. Ti can also be used to pinpoint properties that are likely to have serious hazards. Below are some of the other key benefits:

• Assists policy makers and managers to understand the ‘make-up’ and changes in housing stock across a borough and how it relates to other policy areas, such as anti-social behaviour
• Creates a solid evidence base for council policy and legal interventions, such as discretionary property licensing schemes
• More effective use of scarce staff resources to proactively target non-compliant rented properties, including unlicensed properties
• Provides a vital insight into landlord behaviour and helps prevent crime and fraud, such as Council Tax evasion

How is it delivered

• Ti is developed in partnership with council teams to ensure the maximum benefits of Ti are delivered
• To ensure compliance with GDPR, no sensitive data is taken away and all analysis can be completed within the council’s digital environment.

If you have any questions about Ti please contact Russell Moffatt russell.moffatt@metastreet.co.uk.