Category: Data Stories

Teaching, Learning, and Achievement: Are Course Evaluations Valid Measures of Instructional Quality at the University of Oregon?

Presenter(s): Emily Wu − Economics, Mathematics And Computer Science

Faculty Mentor(s): Bill Harbaugh

Data Story Session 1CS

Research Area: Economics

This study explores the legitimacy of the use of Student Evaluations of Teaching (SETs) as a measure of teaching quality. To do so, we seek to answer two questions surrounding the creation and implications of SETs. Using data from the University of Oregon (UO) we first analyze the influence of a variety of factors commonly hypothesized to bias SET scores. Second,
we investigate the relationship between SET scores and future student achievement. We find that a many of these factors influence SET scores, and that SET scores for a class are not a useful measure for predicting how well students will do in future classes. These findings suggest that SET scores are not a valid measure of teaching quality at the UO.

A Decision Tree Model to Predict Cervical Cancer Screening

Presenter(s): Seth Temple − Mathematics

Faculty Mentor(s): Stephen Fickas

Data Story

Research Area: Natural/Physical Science

I develop a decision tree model to predict if a female patient will be screened for cervical cancer. This project interests me because I want to apply machine learning to improve the health care system. I access the data from the website Kaggle. I use the pandas package to clean the data, and I wrangle some numerical columns with k-means clustering. Graphs will be produced by matplotlib. This project gives me practice in modeling with binary variables. As I plan to enter the actuarial field, this skill set is needed in building fraud and inspection models.

Visualizing Assault Reports in Seattle, Washington

Presenter(s): Lillie Scarth − Spatial Data Science & Technology, Anthropology

Faculty Mentor(s): Joanna Merson

Data Story Session 1CS

Research Area: Social Science

The Seattle Police Department maintains robust, qualitative, open data on criminal activity in the city. While working with previous versions of these data, high rates of assault and abduction among Asian-American women in Southeastern Seattle were identified. These observations corresponded with investigations inside a sex trafficking ring. The goal is to continue this exploration with the most recent version of these data and explore interactive, animated displays using mapping APIs or R Studio. The intent of this exploration is to recognize and display that safety is an intersectional issue.