The spread of COVID-19 and the public’s desire for information has sparked the creation of open-source data sets and visualizations, paving the way for a discipline we’ll introduce as pandemic analytics. Analytics is the aggregation and examination of data from many sources to derive insights, and when used to study and fight global outbreaks, pandemic analytics is a modern way to combat a problem as old as humanity itself: the proliferation of disease.
Here are three ways pandemic analytics are helping us get through the COVID-19 crisis:
To Craft the Right Response
In the early 1850s, as London battled a rampant rise in the number of cholera cases, John Snow – the founder of modern epidemiology – noticed cluster patterns of cholera cases around water pumps. This discovery allowed scientists to leverage data to combat pandemics for the first time, driving their efforts towards quantifying the risk, identifying the enemy, and devising an appropriate response strategy.
That early flash of genius has since advanced, and 170 years of cumulative intelligence has proven that early interventions disrupt the spread of disease. However analysis, decisioning and its subsequent intervention can only be effective when it first takes into consideration all accessible/meaningful data points.
At Sheba Medical Center in Israel, healthcare administrators are using data-driven forecasting to optimize allocation of personnel and resources in advance of potential local outbreaks. These solutions are powered by machine learning algorithms that offer predictive insights based on all accessible data about the spread of the disease, such as confirmed cases, deaths, test results, contact tracing, population density, demographics, migration flow, availability of medical resources, and pharma stockpiles.
Viral spread has a small silver lining: the exponential creation of new data which we can learn from and act upon.
To See the Unseeable
The accessibility of data from trusted sources has led to unprecedented sharing of visualizations and messages to educate the public. Take for example the dynamic world map created by Johns Hopkins’ Center for Systems Science and Engineering, and these brilliantly simple yet enlightening animations from the Washington Post. Such visualizations are quickly teaching the public about how viruses spread, and which individual actions can help or hinder that spread. The democratization of data and analytics tools, combined with mass ability to share information via the internet, has allowed us to witness the impressive power of data used for good.
In recent months, companies have brought pandemic data gathering in-house to develop their own proprietary intelligence.
HCL realized early in the outbreak that it would need its own command center dedicated to COVID-19 response. Coordinated by senior leadership, it gives HCL data scientists the autonomy to develop creative and pragmatic insights for more informed decisioning.
With the goal of enabling leadership to respond quickly throughout the development of the COVID situation, we employed techniques such as statistics, control theory, simulation modelling and Natural Language Processing (NLP). For simplicity, we’ll categorize our approach under the Track & Respond umbrella:
- TRACK the situation quantitatively and qualitatively to understand its magnitude.
- Perform topic modeling in real-time across thousands of publications from international health agencies and credible news outlets; automate the extraction of quantifiable trends (alerts) and actionable information relevant to a manager’s role & responsibility.
- Create forecasting which will directionally track and predict when regions critical to HCL and its customers will reach peak infection, and conversely, a rise in recovery rate.
- RESPOND using a mathematical model of the situation as a proxy for the actual pandemic.
- Create a multi-dimensional simulation model using robust and contextual variables to produce a meaningful prediction customized to the leader using it.
To Diagnose, Treat, and Cure
On December 21, 2019, an AI system operated by a Toronto-based startup called BlueDot detected the earliest anomalies relating to what was then considered a mysterious pneumonia strain in Wuhan. The AI system accessed over one million articles in 65 languages to detect a similarity to the 2003 SARS outbreak. It was only nine days later that the WHO alerted the wider public about the emergence of this new danger.
It’s quite possible that AI in conjunction with medical researchers can help reduce drug development timelines to mere months or weeks.
Where We Go from Here
It is important to remember that technology is nothing but the cumulative innovation of humanity over time, and in technology we have the tools necessary to help us survive and protect ourselves.