Data Science Central Weekly Digest, 23 Nov 2020

From the Editor’s Desk

American Thanksgiving will be … weird this week. The Covid-19 pandemic has entered its second wave with a vengeance, as worldwide there will be 60 million cases and 1.5 million deaths, with a quarter-million of those deaths in the United States. Families will likely be gathering together only over videoconferencing software, which ironically will result in fewer turkeys being consumed. People are still traveling, but air traffic is likely to be way down this year compared to last.

It’s worth reflecting on both the power and the danger of data science and machine learning. The power should be obvious, but let me state it: Within nine months, in a world-wide venture across thousands of different research groups, we’re seeing vaccine solutions and treatments appear in an amazingly short period of time. This has happened primarily because these groups communicated their data and findings with one another.

This is important to remember. Hoarding data does a service to no one. Without adequate data sharing, too much time is spent trying to pull together critical data in the first place, too little oversight exists on the data to make sure it is both valid and usable, and all too often this translates into too much duplication in dead-end research.

At the same time, there is also too much danger coming from the politicization of data. One of the responsibilities of data scientists is to speak up for the science behind that data, to make it clear that good data collection and analysis benefits everyone in the long run.