Want more Insights from SAS? Subscribe to our Insights newsletter. Or check back often to get more insights on the topics you care about, including analytics, big data, data management, marketing, and risk & fraud.
Using Analytics to Make Cities Smarter
By: Tina Schweihofer, Pre-Sales Manager, Data Sciences SAS Canada
The Regional Municipality of York stretches north from Toronto to Lake Simcoe and includes many hectares of protected Greenbelt.
York Region consists of nine local towns and cities, and provides a variety of programs and services to almost 1.2 million residents, 50,000 businesses and 580,000 employees.
Like other organizations, York Region collects massive amounts of data—York Region’s Environmental Services Department, collects approximately, 23,000 data points per minute, or 12 billion per year. Typically, this data has been used for reporting, regulatory compliance, performance planning and the like—in other words, what has happened. But there’s potential to predict what will happen given some innovative approaches to processes. Predictive analytics supplies numbers to help make the case for such innovation to higher levels of management.
In many ways, government sector organizations are risk-averse; big changes aren’t taken lightly. But programs of small changes add up, especially in terms of York Region’s power consumption.
York Region’s Environmental Services Department (ESD), is responsible for ensuring the safety of drinking and wastewater to all nine local municipalities, making them one of the biggest consumers of energy due to their pumping facilities in the Region—and a great place to start a pilot project in predictive analytics.
While ESD had an excellent data team on hand, the Department is quickly realizing that a different skill set are needed to make this proactive shift in their business. ESD partnered with the Masters of Analytics program at York University’s Schulich School of Business to identify a small project that could demonstrate results without compromising quality and service delivery standards—never an option.
The pilot was a four-step process:
- Identifying the ideal operation for the pilot, in this case, the sewage pumping station in the regional seat of Newmarket, with a population of about 85,000.
- Engaging York Region’s Water and Wastewater Operations Team to collect the data. The pump operators have unique insight into the challenges the project might face.
- Having a data scientist create models and apply them to the data using SAS products to determine the best approach.
- Share the results and get buy-in from Operations. This is critical to the success of the operation. Once Operations is onside, formalize the approach and implement the project.
Predictive analytics uses techniques like data mining, statistics and modeling to forecast from current data what future outcomes might be under different simulated scenarios. In the case of the Newmarket pumping station, a time series analysis revealed changes in efficiency when pumps operated together rather than individually, and created a scenario that maximized pumping efficiency while lowering power draw by almost 35,000 kilowatt hours annually—a $12,000 saving. Data scientists also modeled seasonal impacts. Importantly, was Schulich program director Murat Kristal’s ability to explain the results in language that could be understood by everyone from the technical teams to senior management.
Over the course of the last three years, York Region followed up with projects at the Aurora Pumping Station and Pressure District 7 (PD7) in Maple. The Region has installed more data loggers at the Newmarket and Aurora locations to further isolate power savings from the pumps from other current draw in the facility like lighting and electrical outlets. Staff at the Maple site of PD7 started operating optimized combinations in the fall of 2016, and formally implemented the system in January 2017. Energy savings are estimated to be between six and 10 per cent per year.
Pilot projects are for learning, and the department has learned a lot already from this work. Four lessons in particular stood out:
- Plan to measure. Ensure proper measurement tools are in place before you start. This allows you to accurately benchmark—note that York Region had to revisit its data logging equipment on two projects to separate pump savings from other electrical use. Take some time to determine what data you need and what story it is supposed to tell.
- Communication. Talk early and talk often with front-line staff and management. Make sure the right people are involved. Kristal had to work from engineering drawings of unknown vintage, and he’s not a mechanical engineer by trade. Having that person with the needed expertise in the room can help get the project moving—and steer when necessary. In the case of Newmarket and Aurora, some capital upgrades had been made specifically to reduce energy consumption. Having that kind of information in hand at the beginning of the project helps guide measurement efforts.
- Plan to commit. Identify roles and responsibilities for the project—who is accountable for what? This is one reason getting that original buy-in from all staff is so critical. Get everyone involved early. Hear them out. Give a lot of thought to the concerns they raise and act on them if need be. And have regular with staff and management to see if the project or process could use some tweaking.
- Competing for talent. This is an ongoing challenge—finding candidates with the skill set to perform in a predictive analytics environment. It’s a difficult skill set to gain, the market is very tight, and public sector organizations don’t have comparable salaries to offer prospects that private sector or consulting organizations do. The Region has been reaching out to schools like Schulich to impress upon students the scope of the opportunity in the public sector; projects like the pumping station pilots barely scratch the surface of the opportunity in the sector.
For more on predictive analytics, the challenges and successes of York Region, followed by an informative Q&A session, watch our webinar.
Tina Schweihofer is Pre-Sales Manager, Data Sciences, SAS Canada. She is passionate about helping people understand how high performance analytics, coupled with the right data strategy can deliver real business benefits. Tina leads a talented data sciences team that helps organizations across industries apply analytics to solve unique business problems using SAS. Tina can be reach at firstname.lastname@example.org