Canada was well-represented in the first-ever SAS global Hackathon as three teams were recognized for their impressive handling of SAS software to solve complex problems.
One of those teams, the Hackanadians, was even featured prominently during SAS chief executive officer Jim Goodnight’s main keynote address Tuesday morning.
Amir Hossini, a senior data scientist from Calgary, together with team members Sarah Aghakhani, Prabhpreet Sidhu, and Heather Friesen – all based in Canada, except for Friesen who had been working in out of the Middle East for the past several months – earned the spotlight thanks to their Traffic Lights for Life project.
In an interview with IT World Canada, Hossini, taken by complete surprise by the news of the win, said the team was tapping into an area of IoT that hadn’t been properly explored: Audio-based intersection management for emergency vehicle prioritization.
Helping streetlights listen for EMS vehicles
Nearly 70 per cent of the world’s population is projected to live in cities by 2050, leading to a reshaping of the way we build urban areas. Governments know the future of large cities will be directly tied to the digital infrastructure that supports it. IDC says that the top 100 cities investing in smart initiatives in 2019 represented nearly 30 per cent of global spending. The growth is increasingly visible in smaller cities where smart city spending is less than $1 million.
While performing research into their project, the Hackanadians squad discovered that collisions are the second leading cause of U.S. firefighter deaths, and hundreds die each year from collisions with emergency vehicles. Many of these accidents occur at intersections.
“Smart cities are a big initiative for governments…and one thing that has been discussed but hasn’t had a solution to is applying AI in traffic and intersection management,” Hossini explained.
Current solutions to this issue comes in the form of expensive transponders that need to be attached to individual vehicles. The range on most of these sensors, Hossini added, is questionable.
To fill the gap in this area, the four-person team created a cost-effective system that uses artificial intelligence and deep learning to turn traffic lights into listening devices that help guide EMS vehicles through busy intersections. The solution uses sensors to pick up sounds, deep learning for decision making and the ability to send commands to traffic lights.
According to a helpful video created by the team, the audio signals are sensed based on predetermined time intervals digitized by Raspberry Pi and transferred to Azure IoT hub and Event hub. From there, they’re transferred to Jupyter to undergo the deep learning phase and eventually, the system spits out commands for the traffic light sensor.
Hackanadians’ AI model correctly distinguished between emergency and non-emergency sounds and triggered the Raspberry Pi sensor on the traffic light accordingly.
Related:
SAS says several teams will earn an invitation to partner with the company to commercialize their ideas. Winners will also get to hold on to their complimentary access to SAS Viya on Azure, which comes with expert guidance from SAS engineers.
When asked what the next steps are for the team, Hossini said he had no idea. But that was a good problem to have, he added.
“I’m overwhelmed,” he said. “I can’t say we’ve figured out what’s next. We thought it was a good idea to participate, but we never thought of winning, especially when we looked at all the other capable teams.”
Channel partner flexes creative muscles
A small team from SAS channel partner Pinnacle Solutions came up with the greatest team name imaginable and combined IoT sensors in a smart shoe insole with SAS AI and machine learning. The end game: Create a risk model to analyze a patient’s movement and create a risk score for the loss of balance and falling.
It was an ambitious idea, and the Red Hot Chili Steppers had only one week to build it.
“We didn’t want to bite off more than we could chew,” said D.J. Penix, president and CEO of Pinnacle Solutions and Red Hot Chili Stepper team member.
With the help of Ontario Tech University, the team cleverly built upon the Berg Balance Assessment with AI and machine learning to create models that analyze a person’s distinct movements and help doctors, insurers, or even athletes monitor health in a completely different way.
The new system features can visualize including data from the BIO_SOLE and other IoT sensors that can provide gait and motion analysis, balance assessments, and categorization of various tasks such as walking, standing, running, or sitting as examples.
Once the modelling process has been developed, Penix explained that the application can be extended to numerous other health indications, including weight management, detection of diseases such as diabetes, pain management, substance abuse, general health monitoring, and more.
“Not a single team member was able to meet face to face,” he noted. Many baseline models for the various body movements were developed by putting family members through different motions and tracking them with a web camera.
“We did a lot of this with out-of-the-box stuff with SAS software,” Penix said. “And we didn’t have to write a lot of code. It was mostly drag and drop.”
The final featured Canadian team during this year’s Hackathon event was Butterfly Data. The SAS partner used the new conversational AI functionality of SAS Viya to build a chatbot to deliver cancer lifestyle advice for patients and their families and friends.
The system provides statistical data from across Canada, nutritional information, directions to local health care services, and even links to trustworthy news websites.
Software for everyone
The topic of accessibility and ease of use was front and centre throughout the week. During a break in the summit action, SAS chief information officer Jay Upchurch spoke with IT World Canada and said large enterprises are quite familiar with their solutions and are well-versed in conversations about SAS Viya, for example.
But there’s a swell of interest from smaller firms now, too, he noted.
“On the SMB and mid-market side, we are seeing new entrants come in every day,” he said.
In recent years, many SAS partners have gotten good at servicing the needs of SMBs that are often looking for quick results with their collected data. Other analytics solutions, according to Upchurch, often “put the burden on the SMB” to do the heavy lifting and create something out of nothing with the data.
“SMBs want that immediate time to value,” he said.
There’s a big push across the company to no longer be the industry’s best-kept secret.
“We want to make sure that the industry sees SAS embracing the market in an open and extendable way,” Upchurch said. “We don’t want any barriers to the adoption of our software.”