Across the world, we’ve empowered a select few policy makers and organizational leaders to make COVID19 safety decisions for us. To the extent that we follow their guidance – in going to a restaurant, wearing a mask, keeping our distance – our lives, and those of our loved ones, are in their hands.
If you are one of these decision makers, as a COVID10 survivor, I’m grateful you’re out there.
But I know your job is extremely challenging. As a data scientist I’ve observed a systematic gap between the data stack and decision makers in many domains, and COVID19 is no exception.
The bottom line: most of us are thinking about COVID19 data wrong. But there’s an emerging consensus, which I’ve attempted to summarize in the table below. I’ll be covering these topics in greater depth moving forward, including sharing with you how leading data scientists are shifting to this new approach.
|Old belief||New Reality|
|COVID19 data is both necessary and also sufficient.||COVID19 data is necessary, but is in many situations not sufficient, to make good decisions in complex environments.|
|We can relax, the vaccine is on its way.||According ourworldindata, over 17,000 people died worldwide on November 30, 2020. That’s over five times the number of people who died in 9/11 every single day. Or, by my calculation, over 41 times the rate of deaths in the Viet Nam war. This means that the time between now and a widespread vaccine will be one of unprecedented illness. We cannot become complacent.|
AND even when we’re past COVID19, our buildings are still unhealthy; 650,000 people die of flu each year; the most recent example of the southern hemisphere having largely skipped flue season this year, with considerably fewer deaths.
|Scientists are expected to both answer questions as well as to make recommendations.||We recognize that making good decisions is a fundamentally different discipline than what science does.|
|We depend on a government expert or panel to make decisions for us.||We follow government rules, but technology now allows us to understand for ourselves the consequences of our decisions.|
|Decisions and guidance is “one size fits all”.||We recognize that “one size fits all” often means that minority groups are disadvantaged.|
|Safety decision rationale is opaque.||Safety decision rationale is transparent.|
|Because of the complexity of the situation, the system can be “gamed” and “grifted” with disinformation.||Because decisions are transparent, we’re more likely to be able to catch hidden agendas.|
|We rely on words and graphs to understand the situation.||We leverage the brain’s “supercomputer”: visual, spatial, and motor thinking, to solve these hard problems.|
|Our best practice is to share what we learn by doing different experiments within different settings (e.g. schools, protests, religious meetings).||Our best practice is to run simulations of the impact of our safety actions within our particular circumstances, and to let artificial people get sick and die so that real ones don’t have to.|
|We resolve conflict through arguments and verbal text and discussion.||We resolve conflict by visually aligning our mental models.|
|We believe that people are good at assimilating data about COVID19 and making good decisions.||We believe that, in complex, invisible, exponential environments, humans make their best decisions when they allow advanced technology (simulation, AI, agent-based modeling, complex systems modeling, and more) to help them.|
|COVID decisions are overwhelming and all we can do to ensure safety is to follow guidelines given to us by experts.||COVID decisions can be understood, and we must hold experts and technology accountable for ensuring we can understand it. Our lives are at stake!|
|Modeling is limited to answering questions about facts, RCTs, and predictions.||The most important modeling is about actions and consequences (which typically uses fact and prediction modeling as an input).|
|We use advanced technology in a limited capacity.||We recognize that advanced technology has created a revolution in personalization: medicine, marketing, and more, and that we must leverage this power.|