From June 3 to June 7, 2019, a dozen graduate students participated in the course “Fundamentals of Ocean/Atmosphere Data Analysis” held at Oslo’s Science Park. Students from four countries took part, with six from Norway, four from Germany, and one each from Ireland and Sweden. The instructor, Jonathan Lilly, is an oceanographer who specializes in data analysis methodologies. The course ran four to five hours each day, with a mixture of lectures, group exercises, and solo lab assignments.
Students were introduced to essential tools for analyzing any type of dataset from oceanography, atmospheric science, or climate. The centerpiece analysis method, called “distributional data analysis”, is a simple yet powerful method for delving into a potentially large, multivariate dataset by examining its statistics in two-dimensional slices. Elementary statistics for univariate or bivariate (e.g. velocity) datasets, simple filtering, data organization and manipulation techniques in Matlab, and data visualization strategies were all addressed. The course also provided innovative training in the mental factors of curiosity, imagination, and objectivity that are essential for researchers. Students applied techniques to their own datasets, and learned further through homework problems and group exercises. At the end of the course, each student submitted a final project summarizing what they had learned in application to their own dataset.
Student feedback after the course was highly positive. The students’ overall rating of the course had a mean of 9.15 and a median and mode of 9 out of 10. The students’ likelihood of recommending the course to others had a mean of 9.0 and a median and mode of 10 out of 10. The interest in taking a follow-up course had a mean of 9.8 and a median and mode of 10 out of 10. Written and oral comments indicated that the students found the course material to be highly useful and inspiring, and also not overlapping with other courses they had taken.
Text: Jonathan Lilly