Home About ATSIP Program Awards Venue Sponsors/Exhibitors

Using Machine Learning to Make Analytics Smarter

Abstract:  While big data analytics, machine learning, and artificial intelligence are often applied to expansive datasets directly to extract meaning from raw data, seldom (if ever) are those valuable methods brought to bear on the user-interactive tools that practitioners use to view and analyze that data. Recognizing that expert experience and intuition are valuable assets in the pursuit of truth in the data, machine learning analysis of user’s preferences and areas of focus in the system can yield information that informs the creation of a system to recommend other facets of the data that may be interesting to the user, based on what they’ve already looked at. This “personal shopper” for data analysis operates similarly to the way Netflix and Amazon recommend movies and products to you based on what you’ve watched or purchased. It also provides a repeatable method of quickly introducing new users to an unfamiliar dataset, and creates opportunities for experienced users to find items that are of interest that they may not have been aware of otherwise.  

As with any large dataset, it takes time to get familiar enough to start to draw meaningful conclusions, especially as the data morphs and changes. Using the core framework provided by the ADVANCE-family of analytics dashboards (which include crash, citation, and other traffic safety datasets), the goal is to minimize the time between the formulation of the question in the analyst’s mind to finding the answer to that question. Using machine learning to classify the analysis pattern of a given session, the system seeks to provide recommendations of other (possibly even unknown) data points that are relevant to the user’s goal. This method, which is dataset agnostic, takes a holistic view of the activities in order to act as a helpful guide while supporting the decision-based activities that are led by the data.

Jeremy Pate - Center for Advanced Public Safety, The University of Alabama



Session Material
Quick Links


Chris Osbourn


fb_logo   linkedin