Machine Learning on your Street: Bringing AI to your neighbourhood
By Lee Ann Dietz
Lee Ann Dietz, Analytics Evangelist and Strategic Marketing Executive at analytics company SAS, shares how governments can implement AI in their analytics strategies.
Property valuation is a critical function for communities that levy a property tax. Properties need to be assessed fairly so owners are paying their fair share − not too much and not too little − of the total property tax. In most U.S. states, as well as many countries around the world, property tax is the greatest source of revenue for a city, county, or local region. For taxpayers, it can be one of the biggest expenses of owning a property. In all communities, but especially those going through economic expansion or recession, the property values change over time, sometimes in ways that shock property owners.
Local governments are challenged to do more with less, be more responsive to citizens, act in a more timely manner and improve transparency. This is especially true in property assessment because it is one of the few contacts that a government has with all property owners. A specific challenge in this arena includes maintaining accurate measures of every factor on every property. Another challenge is the potential for subjective bias when community appraisers tweak assessments to try to stay in line with changing property values. Finally, in some areas, the rate of owner appeals can be as much as half of the entire number of properties. Government agencies do not have the resources to successfully rebut every erroneous appeal, leading to lower than anticipated revenues. Enter Artificial Intelligence.
Using automated machine learning algorithms that incorporate historical and new property sales transaction data every night, Wake County can recalculate property values as market conditions change. Using a traditional three or four year reappraisal cycle, most governments cannot react to market changes when large businesses move in or out, when weather conditions transform a neighborhood or when there is a shift in the renovation rates for existing houses. The Wake County model can quickly and confidently build a report of significant factors that influence value. A bonus is that the machine learning algorithms provide critical information requested by property appraisal agencies. This information is needed by local government agencies to ensure that bias is not creeping into the model and thus, the property valuation for its citizens. And, Wake County can get this information every day.
Having this validated information available for each property every day means that a local government can begin to project revenues for its property base and be confident that the infrastructure projects that it initiates will have the long-term sustainable revenue to see those projects to fruition. As local governments become more sophisticated and review the factors that influence value, it can optimize infrastructure that is relevant to property values so that property values rise. It’s a win-win for both governments and their citizens.
Citizens have a tremendous interest in governmental transparency and encouraging objective measures to promote fairness and equality. Artificial Intelligence can often be, in my opinion, unfairly labeled as opaque to end users. The machine learning algorithms used by Wake County have the benefit of transparency because they identify the significant variables that influence value. For instance, the models show that finishing a basement adds considerable value while having a second fireplace does not. It’s the same micro-segmentation approach that leading edge companies have been using to sell the newest book on Amazon. An agency could even provide those useful insights to help citizens improve their home’s value.
The revenue derived from property taxes is a simple equation of property value multiplied by tax rate across all properties in the jurisdiction. You wouldn’t think that applying machine learning algorithms would be a game changer. But, in terms of transparency, responsiveness, speed and objectivity, the introduction of this artificial intelligence methodology is exactly that. Just think about applying similar machine learning techniques across other tax agencies. To appropriate another Ben Franklin saying, “to succeed, jump as quickly at opportunities as you do at conclusions.” The opportunity to apply machine learning algorithms in public sector agencies is here: Jump in.
Learn the SAS approach to AI and how government agencies can implement it into their analytics strategies. Download the white paper by filling out the form below:
Lee Ann Dietz supports public sector agencies and transportation companies around the world, helping them solve entrenched problems and innovate new revenue sources by applying data management and analytics technologies. Her focus includes supporting SAS Internet of Things (IoT) and advanced analytics/Artificial Intelligence (AI) offerings around Smart Cities.
In her current role at SAS, she works with cities and transportation companies around the world, providing thought leadership across a variety of Smart City and transportation use cases. She also leads SAS’ partnership efforts with the International Transport Forum. Lee Ann has an undergraduate degree in Economics from Stanford University and a MBA from the Darden Graduate School of Business at the University of Virginia.