Overcoming COVID-19 with Data Science
“We show our students how to collect COVID-19 data, cleansing the data to build models and patterns to help identify the peaks and flatten the curve.”
Dr. Majed Al-Ghandour
The sudden outbreak of the coronavirus pandemic exposed dangerous vulnerabilities in healthcare systems and supply chains around the world. With protective equipment shortages, lack of effective treatment, or even a widespread tracking system, it ran unchecked in its earliest stages.
But it also opened a door for data science to find innovative ways to safely manage future outbreaks and maybe even fast-track a vaccine for this mysterious and deadly disease, said Dr. Majed Al-Ghandour, instructor for the UCI Division of Continuing Education’s Introduction to Data Science course.
Even now, a wealth of big data is being mined and analyzed by artificial intelligence and machine learning, to find the safest and most effective path forward to recovery.
“For one thing, artificial intelligence is being used to show and visualize trends with the spread of the disease, then predict COVID-19 hot spots, tracking and fighting the pandemic with good data models,” Al-Ghandour said.
“We need to learn from COVID-19 patients’ data the best algorithms to use, based on their medical data and a few other things. Using this information, the goal is to train a machine to learn how to analyze COVID-19 data and lead us to innovative solutions.”
Think of data science, AI and ML as a unified weapon in the fight, a black box of sorts, he added. “It’s very basic, like a black closed box. Data in, data out.”
That black box is working overtime to help heal the damage caused by COVID-19 and prevent future pandemics. Far from leading mankind into a “Black Mirror” dystopia, it might be the thing that saves us.
Sharing the data
AI has been especially useful for navigating the delicate path to fully reopening the country, by tracking and forecasting outbreak hot spots and pinpointing ZIP codes with the highest rates of coronavirus cases. Systems using maps and models were developed by several universities and agencies, including Johns Hopkins, the CDC, and the Institute for Health Metrics and Evaluations.
“Counties and states are using AI models to move into the various phases of their reopening plans,” Al-Ghandour said. “Data science analytics can also help predict economic funding shortages to better allocate resources to fight the pandemic. Drones have even been used to patrol public spaces and get medical supplies where they need to go.”
Data science has proven to be a powerful tool in the early stages of developing a COVID-19 vaccine, he added. “Google used algorithms and its computing power to understand the proteins that might make up the virus, then they shared their data with others to help develop treatments.”
Google’s DeepMind division published its AI findings so medical researchers could study the proteins. And BenevolentAI, a company that was already using AI to create drugs to treat serious diseases, is now using its resources to develop new medicines for COVID-19 and investigate existing drugs that might be useful in the fight.
“AI also can be used to accurately diagnose the virus within seconds,” Al-Ghandour said. “A company called BlueDot developed AI algorithm near-real-time insights to help hospitals and airlines prevent outbreaks.”
Tracking the risk
Aside from addressing the need for new treatments, data science tools can help refine tracking systems to trace the movements of infected or potentially infected individuals — a move that has generated controversy over privacy issues, especially in the U.S.
“Contact tracing is a method used to find and follow-up with people who have been in close contact with someone who tested positive for COVID-19,” Al-Ghandour said. “Apple and Google launched new contact-tracing tools using AI and ML, then lent them to public health agencies around the world. These are custom-built apps that run on tablets and smartphones, and they provide a cost-effective and secure way to fight viruses or diseases on national levels.”
Tracking travel activity is an especially useful tool for minimizing cross-border and cross-continent spread of the virus. “Smartphone location-tracking data were made available from numerous sources to monitor travel to reduce COVID-19 hotspots. Closing travel from China and Europe relatively early is a good example.”
Mastering the machines
None of these analytic tools would be effective, of course, without well-trained professionals to master them. The UCI DCE Data Science certificate program aims to open the door to this technology and train the next generation of strategic leaders and managers to be proactive for future emergencies.
The online program provides deep background on data science basics along with hands-on experiential training that addresses the latest applications being used to track COVID-19 cases. Students analyze potential business impact and project how best to distribute resources, for instance.
“We show our students how to collect COVID-19 data, cleansing the data to build models and patterns to help identify the peaks and flatten the curve,” Al-Ghandour said. “We also use predictive analytics to optimize business performance at a variety of levels impacted by COVID-19.”
The program uses the latest visualization software tools to create a dashboard for tracking coronavirus cases, creating maps and charts that break down the latest statistics: new cases, seven-day rolling averages, completed tests and number of patients currently hospitalized. A Hospital Resource Use feature allows students to project and analyze the number of ventilators and hospital and ICU beds that will be needed.
“Coronavirus will cause us to rethink the traditional way of using data science, ML and AI for operating hospitals and our healthcare system,” Al-Ghandour said. “We have a unique opportunity to use these analytic tools to truly evaluate our health agencies, medical supply distribution, the technologies we use as well as our policies and procedures.”
Learn more about the Data Science program.