Statistics

21 posts
Challenges of making a reliable Covid-19 model

Fatalities from Covid-19 range from the hundreds of thousands to the millions. Nobody knows for sure. These predictions are based on statistical models, which are based on data, which aren’t consistent and reliable yet. FiveThirtyEight, whose bread and butter is models and forecasts, breaks down the challenges of making a model and why they haven’t provided any. Tags: coronavirus, FiveThirtyEight, modeling

0 0
Johns Hopkins providing course on using epidemiology to understand the Covid-19 numbers

The numbers are fuzzy. You take them at face value, and you end up with fuzzy interpretations. Starting at the end of this month, Johns Hopkins is providing a two-week epidemiology course on understanding these numbers better: This free Teach-Out is for anyone who has been curious about how we identify and measure outbreaks like the COVID-19 epidemic and wants to understand the epidemiology of these infections. The COVID-19 epidemic...

0 0
Communicating a crisis

David Spiegelhalter on communicating a crisis: There are some basic principles, which I learnt from John Krebs, former Chair of the Food Standard Agency, who had to deal with many crises. The first thing is that you should be communicating a lot, consistently and with trusted sources. You have to be open and transparent. You have to say what you do know and then you have to say what you...

0 0
Understanding data and statistics in the medical literature

Jeffrey Leek, Lucy D’Agostino McGowan, and Elizabeth Matsui have a free/ pay-what-you-want book on understanding data and statistics in the medical literature: Whether you are a medical student reading their first journal article or a healthcare professional trying to use the latest research to improve patient care, understanding data and statistics has never been more fundamental for extracting accurate information from the medical literature. This is a high-level, introduction to...

0 0
Possible coronavirus deaths compared against other causes

Based on estimates from public health researcher James Lawler, The Upshot shows the range of coronavirus deaths, given variable infection and fatality rate. Adjust with the sliders and see how the death count (over a year) compares against other major causes of death: Dr. Lawler’s estimate, 480,000 deaths, is higher than the number who die in a year from dementia, emphysema, stroke or diabetes. There are only two causes of...

0 0
Proportions and rates: we are no dupes

Reader Lucia G. sent me this chart, from Ars Technica's FAQ about the coronavirus: She notices something wrong with the axis. The designer took the advice not to make a dual axis, but didn't realize that the two metrics are not measured on the same scale even though both are expressed as percentages. The blue bars, labeled "cases", is a distribution of cases by age group. The sum of the...

0 0
Federal budget scaled to per person dollars

For The Upshot, Alicia Parlapiano and Quoctrung Bui scaled down the federal budget to something more relatable: To better understand how federal spending has changed since Mr. Trump has taken office, we looked at the actual budget amounts for the 2020 fiscal year. We divided them by the U.S. population and sized the numbers proportionally to make their scale easier to visualize. Then we compared the numbers to the actual...

0 0
Billionaire’s spending scaled to your net worth

We hear about billionaires spending millions of dollars on ads, acquisitions, etc. It seems like a ridiculous amount of money, but that’s partially because us common folk think of the millions of dollars in the context of our own net worth. When Jeff Bezos spends a few multiples of what we will never make in a lifetime, it seems like a lot. For The Washington Post, Michelle Ye Hee Lee...

0 0
Data problems in Iowa caucus results

It wasn’t just issues with an app. There appears to be many more problems with the Iowa caucus results. The Upshot broke it down with a closer look at the data: Some of these inconsistencies may prove to be innocuous, and they do not indicate an intentional effort to compromise or rig the result. There is no apparent bias in favor of the leaders Pete Buttigieg or Bernie Sanders, meaning...

0 0
Privacy algorithm could lead to Census undercount of small towns

To increase anonymity in the Census records, the bureau is testing an algorithm that removes real people and inserts imaginary people in various locations. As you can imagine, this carries a set of challenges. Gus Wezerek and David Van Riper for New York Times Opinion ask what effects this could have on small towns. Tags: census, counting, New York Times, privacy

0 0