Data runs the world. With the continued development of data capture tools and the constant entry of older data into software platforms that can leverage it, businesses and governments today find themselves sitting on a veritable goldmine of machine-friendly information. But what can organizations do with this data? As it turns out, quite a bit. And some of the most promising technological advances to emerge from our data-driven world can be seen in the development of predictive analytics.
What are predictive analytics?
Predictive analytics is a form of analytics that leverages a massive wealth of data and sophisticated algorithms to create predictions about the future. Data scientists use deep learning and machine learning algorithms (such as linear/nonlinear regression, neural networks, and decision trees) to create predictive models that help organizations plan for future trends and developments. Predictive analytics can be used in many contexts, from voice-to-text translation to weather forecasts, investment portfolio development, and beyond.
Regulators are no exception to this rule. Thanks to the massive increase in available data regarding federal and state-level regulations, regulators find themselves presented with an opportunity to gain a new level of insight into the effects and outcomes of state policies. In occupational licensing, specifically, we see a unique use-case for predictive analytics to gain insight on recidivism, wage discrepancies, racial disparities, quality of services, and much more.
Gathering data for predictive analytics
In the United States, where each state has its own unique administrative codes, recent developments in data-mining technology have allowed researchers to algorithmically scan through these codes and pull out all relevant information, as opposed to gathering this data by hand. An example of this can be found in the Occupational Licensing RegData (OL RegData) database, which comes from the Mercatus Center’s QuantGov platform. RegData leverages machine learning in an attempt to quickly and accurately capture the extent and complexity of regulations in each state. Its algorithm is trained by searching for sections in state codes that regulate participation in an occupation.
There are advantages and disadvantages to using machine learning to capture and analyze occupational licensing codes. Of course, using an algorithm like RegData allows researchers to save countless hours that would otherwise be spent gathering data by hand. But this comes at a cost. Manual data collection, tedious as it may be, still offers researchers a level of detail that cannot always be achieved with machine learning. When data is collected by hand, researchers can make notes within the dataset that provide context for what may appear to be outliers or errors. Machine learning, as it exists today, cannot always achieve this level of granularity.
The OL RegData counts the number of restrictions, the number of words, and the average sentence length in a state’s occupational licensing code. By combining these three variables, researchers can gain insight into how restrictive or even burdensome a state’s licensing regulations can be. The advantage of manual data collection here is that researchers working through state licensing codes by hand can capture many more variables, including license fees, educational requirements, and more peculiar ones like Indiana’s policy of licensure by endorsement.
How can predictive analytics guide occupational licensing bodies?
With the immense amount of data at hand regarding the intentions and effects of different occupational licensing policies, regulators can make critical decisions to improve quality of practice, reduce racial disparities, or otherwise alter their policies in a way that benefits the public. For example, using predictive analytics to gain insight on complaints can help regulators notice trends in the public’s experience with their field and proactively adjust their policies to account for these trends.
Consider, for example, the current worldwide labor shortage that has arisen because of the COVID-19 pandemic. In a field like healthcare, where the stakes are life and death, using predictive analytics on complaints filed by dissatisfied patients can help regulators understand just how severe the worst consequences of this shortage are. In response, occupational licensing bodies can relax their licensure standards accordingly to bring new workers into their field with more ease and efficiency.
Predictive analytics will only grow stronger as machine learning algorithms improve and as more information is amassed in databases around the world. The technology has a strong influence in almost every part of the private sector, and if businesses are using predictive analytics to improve their products or services based on the insight they provide, regulators have an obligation in their own fields to explore the myriad applications of this technology, which has proven to be one of the data-driven world’s most influential developments.
