Home » Columns, Featured, Stats4Good

Data for Good Highlights at JSM 2019

1 July 2019 1,177 views No Comment
This column is written for those interested in learning about the world of Data for Good, where statistical analysis is dedicated to good causes that benefit our lives, our communities, and our world. If you would like to know more or have ideas for articles, contact David Corliss.

David CorlissWith a PhD in statistical astrophysics, David Corliss leads a data science team at Fiat Chrysler. He serves on the steering committee for the Conference on Statistical Practice and is the founder of Peace-Work, a volunteer cooperative of statisticians and data scientists providing analytic support for charitable groups and applying statistical methods in issue-driven advocacy.

 

The Joint Statistical Meetings is one of the largest gatherings of statisticians in the world, making it one of the most important ways to learn about Data for Good, see presentations on the latest methodologies, get involved with new projects, and discuss the important challenges we face every day. JSM’s theme this year is Statistics: Making an Impact, and Data for Good is one of the best ways to make a huge impact with statistical science.

With so many presentations and other activities at JSM, it’s easy to become overwhelmed. Using the search engine on the JSM website’s main page (it’s in the upper-right corner), you can look for the sessions, subjects, presentations, and authors you want to see. Even if you won’t be attending JSM, searching the presentations is one of the best ways to find out about new developments, get in contact with researchers, and begin using statistics to support the cases and causes that mean the most.

Looking through the opportunities at JSM this year, here are just a few that stand out (but there are many more):

  • On Sunday at 4:00 p.m., the Social Statistics Section is sponsoring a session (#50) with papers on measuring discrimination and its impact. Examples include salary differentials by gender, the impact of voter ID laws, and bias in predictive policing algorithms. These papers are important both for their subject and also for learning methods that can be applied to other areas. For example, the biased predictor variables in Lynne Billard’s paper on salary differences by gender are a problem affecting many types of discrimination.
  • On Monday at 8:30 a.m., you can start your day with Administrative Income Data, Survey Data, and Inequality (#106). Subjects include evaluating the distribution of personal income, income tax noncompliance, and a historical review of President Johnson’s war on poverty. In addition, Bruce D. Meyer and Derek Wu will present on a new way to estimate poverty—something critical today in my view—with Mollie Orshansky’s original poverty line estimates of the 1960s only updated for inflation, not rebuilt from first principles. As a result, longitudinal changes have undermined the degree to which the official numbers reflect the reality of poverty in America today.
  • On Thursday at 10:30 a.m., there will be an important session (#630) on machine learning applications in criminal justice and how to address some of the problems found there. Kristian Lum from Human Rights Data Analysis Group will present an algorithm for removing sensitive data, Richard Berk and Ayya Elzarka from the University of Pennsylvania will speak about fairness in risk forecasting in criminal justice, and Cynthia Rudin from Duke will present on the hot topic of black box models. These applications touch on some of the most important issues in Data for Good today, and the methods to be presented can be used in areas beyond criminal justice.

In addition to topics specific to Data for Good, JSM attendees will want to look for presentation about statistical methodologies needed in our research. For example, there will be a session titled Multiple Imputation Monday at 8:30 a.m. On Tuesday at 10:30 a.m., Andrew Hoegh will present a multiscale spatiotemporal model for forecasting civil unrest. Methodological presentations such as these are essential for learning the latest techniques to be applied in a wide variety of areas.

On Sunday at 4:00 p.m., there will be a session titled Improving Data Collection: Challenges in Survey Practice from the Government Statistics, Social Statistics, and Survey Research Methods sections. This takes place at the same time as the discrimination session, so I’ll be reading the papers in the JSM Proceedings and trying to catch up with some of the authors to discuss their work in person.

Students and early-career statisticians will want to check out new topics. Statistics Without Borders has a session on Sunday at 2:00 p.m. that will introduce you to the work of this leading Data for Good organization. Also, keep in mind that one of the best ways for people to get started in Data for Good is providing statistical analysis for a group to which you already have a connection. Look for presentations in an area in which you are volunteering now and see how you can get involved.

Posters are another great way to meet new people and learn about new frontiers and methodologies in Data for Good. Rajarshi Dey and Andrei Pavelescu’s poster on gerrymandering and Nilesh Shah’s investigation of racial bias in calling one type of penalty in football games sound like good ones to have on your list.

JSM is the biggest event of the year. A little careful planning can go a long way toward turning it into opportunities for Statistics: Making an Impact. Even if you aren’t going to JSM, have a look on the website to find other people and projects in your areas of interest and be sure to look at the JSM Proceedings when they come out. On a personal note, I hope to have the opportunity to meet and talk with many of you this year at JSM! If not, you can write me at davidjcorliss@gmail.com. I am always interested in receiving input, feedback, and ideas for what you would like to see here each month.

1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)
Loading...

Comments are closed.