Our mission at Covid Act Now is to enable leaders to make better, faster decisions that save lives, informed by best available data and modeling. To do this, we feel that it’s important for us to follow the open-source methodology to make sure that our team is benefitting from the best available approaches, software tools, and data produced by the many dozens of engineering and research teams studying COVID.
To that end, we want to share our philosophy to modeling COVID:
With a disease that spreads as fast as COVID, there is no time for ego. If there is a better (open source) model than ours available we want to use it. (In fact, we’ve already done so twice.)
Providing the best available data and tools available means being objective and not letting unfamiliarity or provenance get in the way of improvement.
Given the stakes with COVID, it can be challenging for researchers and decision-makers to tell the unvarnished truth. Constructively interpreting and sharing this data can be hard.
We feel strongly, however, that it is better to err on the side of being open, transparent, and truth-seeking when publishing COVID data and modeling.
Why? Publicly accessible data brings clarity, and clarity speeds up the decision-making loop, which is crucial to the fight of a rapidly spreading disease such as COVID.
We also believe that the public has large a stake in all of this, and deserves a clear, plain understanding of COVID, however uncomfortable.
Combine efforts with others
Whenever possible, we choose to combine our efforts with others.
Our modeling technology started as an internal project. A week-plus later we joined forces with Georgetown infectious disease modelers. Most recently, we’ve partnered with the modeling team at Grand Rounds.
We see this approach of combining efforts with other teams as a strength. We plan to keep doing it. Consolidating expertise and de-duplicating work speeds progress and will ultimately save more lives.
Work in the open
To speed up the research and decision making loop, we believe it’s important that modeling work is done out in the open as much as possible.
We recognize there is risk in publishing work, but feel the benefits (more eyes, more input) outweigh the downside. The more scientists and engineers work in the open, the faster they can learn from one another, and the better they will be able to serve decision-makers and the public with accurate life-saving tools.
To that end, we are following the common open-source methodology of publishing our code, assumptions, and thinking. We will err on the side of transparency, even if it means making mistakes in public. We will learn from our mistakes quickly.
Today there are likely more than a hundred separate teams working to understand and model how COVID propagates. We are committed to ensuring our COVID modeling work is additive to this ecosystem and clarifying, which means being thoughtful about how our work intersects and leverages the work of others’.
We have greatly benefited from the aggregated knowledge of talented epidemiologists, data scientists, engineers, public health researchers, and front-line experts working 24×7 across the world to beat COVID. We hope to support their work in kind to the best of our collective ability.
In the end, beating COVID will be the story of science, leadership, and human ingenuity at its best.