Summer Forecasting

Data updates

Before forecasting, it may be helpful to update DESR and Class List data from MyReports. See the data page for details on how to do this.

Forecasting can only really estimate enrollments for courses we’ve offered in previous term types. Forecasting a list of relevant course from the previous term type is a good start to see if demand has increased. Run these commands to generate a list of courses from last summer and then forecast for them

Rscript cedar.R -f enrl -t 202460 --uel -l lower --college AS -a course --output csv

Rscript cedar.R -f forecast -c enrollments.csv -t 202560

IMPORTANT REMINDER: The regstats reports are good indicators of NEW courses that might enroll well, even for summer, and likely better show impacts of new scholarship rules.

Create a regstats report to see what courses are being flagged (based on threshold filtering for enrollment/drop anomolies) in the term(s) PRIOR to the term you want to forecast for:

Rscript cedar.R -f regstats -t 202510 --output csv

This creates a group of CSV files in the output/csv folder, in case you want to handle them separately, but…

All “flagged” courses (meaning they seem to deviate from their normal enrollments or drops, etc) are combined in all_flagged_courses.csv

You can also specify–output html or –output aspx if you want non-CSV files to share

To forecast for flagged courses (via the CSV file created above):

Rscript cedar.R -f forecast -c all_flagged_courses.csv -t 202560

This creates two CSV files: output/csv/forecasts-short.csv

Use the forecast-report function with filtering to display forecasts of interest (like all summer forecasts), for determining needed sections

  • Rscript cedar.R -f forecast-report -t summer --output csv
  • Rscript cedar.R -f forecast-report -c ENGL 1110 --output html

Regstats also creates a csv file called high_fall_sophs.csv that lists courses that sophomores take in the fall and therefore may be useful to offer in summer

Augment regstats with course reports

Because COURSE REPORTS provide valuable insight into course dynamics over the last few years, they help determine if the algorithmic forecasts/recommendations make sense and identify new courses to offer

Especially useful for investigating courses that haven’t been offered in summer but maybe should be!

Course reports can be created individually:

  • Rscript cedar.R -f course-report -c 'BIOL 1110'

Easier to generate a set of course reports from some of the CSV files created by regstats (in the output/csv folder):

  • Rscript cedar.R -f course-report -c drops.csv

You can generate course reports for courses you’ve already forecasted for:

  • Rscript cedar.R -f course-report -c forecasts

WARNING: This generates probably way more reports than you really need!