What statistics students in natural resources want to learn more about

December 22, 2021
teaching statistics analytics Data science statitics

As the holidays approach, it’s always a great time of year to reflect back and look forward. This post will describe my approach to this as it relates to a class that just wrapped up for the Fall 2021 semester.

I instruct the course NR 5021: Statistics for Agricultural and Natural Resources Professionals at the University of Minnesota. The class has between 25 and 40 graduate students enrolled in a given year, a mix between students studying natural resources, agriculture, and plant sciences. Students are typically in their first year and enrolled in a Master’s or PhD program.

The course presents introductory concepts in statistics with an applied focus. We move from probability concepts all the way through zero-inflated regression, a pace that has worked well for the graduate students. We use R as software and students complete assignments in R Markdown in addition to other quizzes and posts related to statistics concepts.

One of the last things I ask students to complete in the class is a poll inquiring about future topics in statistics they would like to learn about. With a solid foundation of statistical concepts, students generally know what skills they need to apply to their own research and what topics they would like to dive into deeper.

Over the last two classes, 55 students have responded to the poll. Students simply check a box next to a topic if they would like to learn more about it, with no limit on the number of selections. The results are:

What strikes me most about the four topics which ranked highest is that they are not exactly topics in statistics. Graphics in R, exploratory data analysis, reproducible research, and database design are certainly related to statistics, but we often refer to these skills as ones that an analyst performs. In many ways they are the “soft skills” of a statistician and don’t require any knowledge about p-values, hypothesis tests, or assumptions of normality.

It always strikes me as odd when I learn that a faculty member is eager to teach a new class in a “hot topic” like machine learning or Bayesian statistics, yet those topics appeal to a niche segment of students. We should be offering more classes and workshops in topics such as creating graphics and exploring data. Such skills will enable students and professionals to better understand their data and communicate it effectively with others.

For now, I’ll share these results with my colleagues and begin thinking about how to integrate these concepts into other courses.

As a part of the class, I’ve put together the book Statistics in Natural Resources: Applications with R. I welcome your feedback on it!

By Matt Russell. Sign up for my monthly newsletter for in-depth analysis on data and analytics in the forest products industry.

Five R packages every forest analyst should be using

January 16, 2022
analytics Data science forest inventory forestry R R packages

Visualizing forest carbon trends in the last thirty years

November 15, 2021
analytics carbon Data viz forest carbon forest industry

Random forests: a tutorial with forestry data

September 27, 2021
analytics Data science random forests statistics