Editor’s Letter—Vol. 34, No. 2
Dear CHANCE Colleagues,
It has now been over a year since SARS-CoV-2 began spreading across the globe. Researchers now understand the virus better, and vaccines are being distributed—but what do we know about the impacts the pandemic has had on our economy and society? Authors Jon T., Nicholas B., and Thomas Middleton study this question as it pertains to the United Kingdom in the article “Modeling the Economic and Societal Impact of Non-Pharmaceutical Interventions During the COVID-19 Pandemic.”
Keeping with the topic of COVID-19, we look at an application of Benford’s law, a tool that has successfully been used in other applications such as detecting financial fraud and altering digital images. In the article “Benford’s Law and COVID-19 Data,” authors Chase Marchand and Dalton Maahs use it to analyze reported COVID-19 cases.
I’ve always been drawn to the personal stories of the men and women who developed the statistical methods we use for data analysis. In “The Secret Career of Solomon Kullback,” historian Brenda McIntire details the U.S. intelligence career of one of the developers of the Kullback-Leibler divergence. If you are intrigued by this article, you may also enjoy the recent PBS special (“The Codebreaker”) featuring one of Kullback and Leibler’s contemporaries, Elizebeth Friedman.
Societal benefits abound from sharing data and machine learning models built on those data, but data have the potential to be biased, resulting in biased models. What are the possible repercussions? Who should be held accountable? What methods can be employed to avoid data bias? Charna Parkey tackles these questions in “Who is Accountable for Data Bias?“
In the February 2021 issue of CHANCE, we included an article about the connection between road familiarity and traffic accidents. In this issue, authors Adam Palayew, Sam Harper, and James Hanley consider a different angle on the topic of traffic accidents. They note that the study of factors that influence accidents may be complicated by factors such as the season, day of the week, and time of day of the accident. The authors evaluate various methods to minimize the effect of these extraneous factors in the article “Toward reducing the possibility of false positive results in epidemiologic studies of traffic crashes.”
Listed as one of the 25 most-dangerous jobs in the U.S. (based on 2019 data from the U.S. Bureau of Labor Statistics Census of Fatal Occupational Injuries), police work is not for the faint of heart. In “Police Officers Killed in the Line of Duty: A Correspondence Analysis of Circumstances and Time of Day,” Terry Allen investigates whether certain types of police activities are more dangerous at specific times of day.
Moved by the horrific 2018 attack on the Pittsburgh Tree of Life synagogue, Howard Wainer and Richard Feinberg analyze data associated with hate crimes. In the Visual Revelations column article “Looking at Reported Hate Crimes,” they zero in on statistics reported in the state of New Jersey due to its uniqueness in having complete data going as far back as 1990.
In The Big Picture column article “The Shape of Things: Topological Data Analysis,” Nicole Lazar and Hyunnam Ryu demonstrate the use of topological techniques for exploratory data analysis. If you work with complex data, this may be a technique to add to your toolbox.
Escape rooms aren’t just a social activity to enjoy with your friends. They can also be used in the classroom to reinforce new concepts. In the Taking a Chance in the Classroom column article “The Data Science Instructional Escape Room: A Successful Experiment,” authors Valerie Nelson and Jason Crea explain their implementation of an escape room for a data science course.
Finally, I’d like to announce that CHANCE will host a session at the Joint Statistical Meetings in August, entitled “The Stories of CHANCE: Frontiersman, Exoneree, Player, Spy.” Whether the conference is held virtually or in-person, I hope to see you there!
Amanda Peterson-Plunkett