The situation the world finds itself in with the coronavirus is unprecedented. An example from a data science perspective is that the nature of the data available and the approach to testing for the virus is changing as we try to measure its impact. This is a catastrophic scenario for many traditional methods in data science, as there is no foundation – the “ground truth” does not exist.
This lack of context impacts decision making and the types of methods that can be effective.
There is no doubt that the coronavirus is disrupting the economy, supply chains and geopolitics. As it progresses, the pandemic will impact different parts of the world’s integrated value chain in different ways and at different times. Even unintentionally, governments will impact each other, leading to compounding disruption.
This compounding disruption also complicates efforts to use data to make informed decisions.
Learning and evolving
If we draw a parallel with businesses, there is usually a point where there asia mobile number list is enough information to make a decision, but not necessarily enough for it to be a good decision. In these types of scenarios, one must constantly review what needs to be understood in order to improve the decision based on new data. As the data changes, decisions and reactions must be adjusted. It is necessary to use methods that are being developed at the moment, based on actions and reactions, since generally in a crisis, there is no historical data to learn from.
Scientific thinking is largely evident in the current crisis, with examples of briefings referring to what was said yesterday, what is believed to be true and therefore what is being done.
Such a research “base” is one of the hallmarks of good science. We know that the key to predicting on-time payment science is always developing (whether in times of crisis or not). We will continue to learn more about the impact of the coronavirus on the world. Good leaders (government and business) must communicate authentically – for example, by adopting a mindset of sharing the best guidance with what each government has today and continuing to learn to adapt responses based on what is learned together. Testing and learning is essential.
Failure is the first attempt at iteration-based learning.
Ask the right questions!
Data doesn’t necessarily “speak”; it’s the interpretation of the data fans data that does. However, anyone interpreting data will have biases. Different methods and techniques have preconditions. Methods like machine learning often require some training or examples.
Right now, we don’t have examples. We need to make inferences, move forward and draw new conclusions. Data is important, but it’s the preconditions, critical thinking, the questions we ask and how we challenge bias that will help us overcome it.
Ask questions to guide your thinking: What data is needed? What predictions are we trying to make? What is being learned? How does it move us forward?
Finding the truth?
When we ask ourselves about finding the “truth,” in any crisis situation, the impact is more about what one believes at a given time and why. There will be many competing “truths” coming from different perspectives, including data latency, observer perspective, intentional alteration or suppression of data, and other factors. During any crisis, there may not be one absolute truth.
The idea is to focus on what you believe and why, and how those beliefs impact the decisions you make. It’s about using questions to guide your thinking: What data is needed? What predictions are you trying to make? What is being learned? How does it move you forward?
What is the best way to evaluate which data to believe?
- Triangular: obtaining the same data from another source.
- Always ask the question “What do I need to believe in order to believe this number?” (e.g., if a country publishes data on the infection rate, the first thing we need to believe is that the country measures the infection rate the same way we do, that the number is current, that the information has not been altered, etc.).