University of Southern California
AC-1: Academic Courses

Data Overview:
Course Data: A complete list of course offerings for three academic years (AY) were obtained from the USC Registrar’s Office and Office of Institutional Research (AY21, AY22 and AY23). The complete list of course offerings included the course number, title, description, course section numbers, instructors, term (semester and year), enrollment number and the time and day of instruction. The Office of Sustainability (USC OoS) corrected carriage spacing errors from the data retrieval process and excluded courses that were missing informative titles or descriptions, such as dissertation or thesis credit courses, overseas course offerings, capstone project credits, research enrollment credit courses, etc. USC OoS removed courses with 0 enrollment from the course dataset.

Full details on how we excluded courses can be viewed in the R.code here: https://github.com/USC-Office-of-Sustainability/SustainabilityCourseFinder/blob/main/01_cleaning_raw_data/00_raw_usc_data/02_formatting.R

Department Data: Since USC does not have a widely-accepted number of academic departments, for this credit, the "total number of academic departments that offer courses" was determined by finding the number of unique departments in the list of course offerings.

Brief Methodology of Classification Process:
This methodology was originally inspired by the UN SDGs (Sustainable Development Goals) keyword mapping (indexing) process used by Carnegie Mellon, Elsevier Publishing Company, Aurora University, and several other institutions, as well as by the sustainability course dashboard initiated by Carnegie Mellon. USC OoS (data analyst Julie Hopper and PSIP data interns Brian Tinsley and Alison Chen) worked with USC’s Presidential Working Group in Education (PWG-E) to further edit and clean the SDG keyword list (modified from the version used by Carnegie Mellon) and used the Text2SDG package in R to map keywords associated with each of the SDGs with text in the course data files.

Prior to keyword indexing, USC OoS removed all punctuation from the course description and title text and transformed all text to lowercase formatting (to assist with keyword mapping). USC OoS ran context-dependency code across the course list to correct keywords in the course title and description based on their context and the adjacent words. (Example: biological law becomes biological domain to avoid mapping to SDG 16 based on the keyword ‘law’)

A full list of our current keywords and our context dependent replacements can be viewed here:
https://github.com/USC-Office-of-Sustainability/SustainabilityCourseFinder/tree/main/Data

To classify as “Sustainability-Focused”, a course title/description had to include at least one keyword that mapped to an environmental SDG (6, 7, 12, 13, 14, 15 ) AND at least one keyword that mapped to a socioeconomic SDG (1-5, 8-11, 16, 17)*. *Designations of which SDGs were environmental and which were socioeconomic were based on the SDG descriptions and targets: https://sdgs.un.org/goals

To classify a course as “Sustainability-Inclusive” a course title description had to include at least two keyword that mapped to at least two of the 17 UN SDGs. OoS implemented the rule that a keyword in a course title/description could only count once per SDG (no duplicate counts for the same keyword if it appeared more than once in the course title/description).

Although not used in this STARS credit, USC OoS also labeled courses as 'SDG-Related' in the attached spreadsheet if they mapped to at least 1 SDG via at least 1 SDG keyword. All other courses (which did not map to any SDG keywords) were labeled as 'Not-Related' (in relation to sustainability).

In the attached file, one can see what keywords mapped to which SDGs from the course title and description and how the courses were classified as 'Sustainability-Focused', 'Sustainability-Inclusive', 'SDG-Related' or 'Note-Related'. *Only the Sustainability-Focused and Sustainability-Inclusive courses were counted as 'Sustainability Courses for the purposes of this STARS credit.

Course Mapping Examples:

Sustainability-Focused Example:
The course: LAW-655 "Environmental Law - Focus on environmental law policy and practice. This course is a combination of regulatory and private law, with a special emphasis on disputes and regulations involving contamination in soil, water and air" mapped to the SDG keywords: law,regulations,policy,soil and thereby mapped to a socioeconomic SDG (SDG 16): "Peace, Justice and Strong Institutions" and to and environmental SDG-15: "Life on Land". In reality, this course should have also mapped to the environmental SDG- 14 "Life Below Water", but the word 'water' was a difficult keyword to map without causing a lot of mismapping issues.

Sustainability-Inclusive Example:
The course OS-512 "Introduction to Chemical and Physical Oceanography - Principles of physical, chemical, and geological oceanography including discussions of air-sea interaction, biogeochemical cycling and the role of the ocean in modulating climate and atmospheric composition; discussion section will cover formulation of basic calculations that illustrate these principles." maps to SDG 13,14 and 15 with the keywords: (air sea,ocean,sea,atmospheric,geological,oceanography). This course is Sustainability-Inclusive rather than Sustainability-Focused as it does not intersect both socioeconomic and environmental-ecological SDG realms.

USC OoS manually reviewed the resulting mapped course list continuously while developing the methodology to achieve better accuracy before achieving the final code (cleaning, context dependency, mapping and classification) and keyword list. USC OoS additionally manually corrected any courses that were automated as 'Sustainability-Focused' but were in fact 'Sustainability-Inclusive', and similarly corrected any courses that were automated as 'Sustainability-Inclusive', 'SDG-Related' or 'Not-Related' but were in fact 'Sustainability-Focused'. USC OoS did not manually correct any other courses to the Sustainability-Inclusive category as the current approach was very conservative on those numbers. However, USC OoS recognizes that there were likely many courses labeled as 'Not-Related' or 'SDG-Related' that should have been labeled as 'Sustainability-Inclusive' but were not, as a consequence of the current conservative methodology. USC OoS will improve this methodology through time with increasing internal and external feedback and collaboration.

All code and datasets can be viewed here: https://github.com/USC-Office-of-Sustainability/SustainabilityCourseFinder/tree/main

Each offering or section of a course was counted as an individual course throughout all calculations (numerator and denominator).

All data and code are public here: https://github.com/USC-Office-of-Sustainability/SustainabilityCourseFinder/tree/main

See attached keyword list used to map the course titles and descriptions to the 17 UN SDGs. We thank Carnegie Mellon for sharing their original list with us, which we modified substantially (see notes in the methods description).

The Sustainability Course Finder dashboard that shows the SDG mapping process can be viewed here: https://usc-sustainability.shinyapps.io/Sustainability-Course-Finder/

Related News Article:
https://docs.google.com/document/d/1LRJ9M4YG9h1RzvqKhaIZnG0A6w4UJ85Vqn4-xqL4-kE/edit?usp=sharing