It’s a New Day in Data: Predictive Analytics is Front & Center at Forum


I was proud to be part of the Omaha Predictive Analytics Forum hosted by HDR on June 16, an event designed to encourage knowledge sharing among area organizations that are using predictive analytics to improve their business. Analytics experts from several large, locally-based firms were represented.

Attendees offered their experiences about how predictive analytics teams can share tools and data for a variety of consumers. We also explored the level of training needed for a data scientist in the areas of statistics, machine learning (area of study surrounding mathematics and computer science) and mathematics.

Four of us from the Predictive Analytics team presented on different aspects of how we’re using predictive analytics to inform the design process — in addition to me, Branden Collingsworth, Matthew Leick Macari and Tyler Oberembt also presented. It’s also exciting to be part of this “new day in data” with all of the analytics and processes involved to make it dynamic and relevant for our clients.

Aggregating Public Data

Matt shared how we store and structure data for analysis. Using data gathered by the US Census through the American Community Survey API, and a proprietary data transformation toolkit, we’re loading data to a geospatially-aware data warehouse, down to the block group level–a geographic unit that comprises anywhere from 600 to 3,000 people. Marrying variables such as gender, age, mobility and language fluency at this level allows us to identify areas that need increased access to services or resources. With this information, we build a complete picture of the community on a granular level.

Location Analysis by Site Access

Explaining how we use spatial analysis for health care planning, Tyler illustrated how important it is to consider the time involved in commuting to and from a hospital when developing strategies for planning for a new hospital. Drive-time analysis is one technique we use to illustrate a birds-eye-view of patient accessibility. This analysis reflects real-time traffic conditions and can be calculated for any time of the day. Drive time analysis is also helpful in determining how close the hospital is to the patient population and even compares distances to competing providers. The map above depicts three different hospitals, all within a 10-minute, 20-minute, or 30-minute drive-time radius.

Increasing Collaboration with Spatial Design

I presented a case study involving an interesting and challenging project we’re working on for an academic medical center. Increasing inter-departmental collaboration is often a primary goal for academic medical centers. Our integrated team, with experts from workplace design, research computational design and predictive analytics, applied innovative methods in planning and design to help uncover paths to increase collaboration. In cooperation with the client, HDR analyzed the relationship between space, social networks and opportunities for collaboration — to inform the design project.

The relationship between walking distance and collaboration is a key metric for the design. After studying walking distances between departments, we were able to predict when collaboration between departments would occur, which allowed our design to impact the organization’s collaboration goals. Building performance (e.g. energy consumption, lighting etc.) the distance metric, spatial analysis and cost modeling are combined to illustrate an optimization problem.

In this case, we integrated design and data to quickly generate solutions based on departmental planning and room-by-room adjacencies, measuring and scoring the design options in real time. Our process is built on a foundation of research, analytics and design optimization to help position our clients to achieve the highest levels of measurable success.

The Future of Predictive Analytics

As the forum came to a close, Branden led a lively discussion that focused on topics such as the future of data science, the role of new technologies and the rise of the “citizen data scientist” (people who are not formally trained in the data science field, but who have access to sophisticated analytics tools).

We were thrilled to host this month’s Omaha Predictive Analytics Forum, share our experiences and learn from this community of analytics professionals. We’re also excited about the forum’s success and how it elevates the art and practice of predictive analytics. Just a guess (not based on any hard data) but we predict there will be more of these forums in the future.