Group project by:
- Andrew Carl
- Charles Casper - https://github.com/ccaspar
- Youn Hee Pernling Frödin - https://github.com/yhpf
Data was collected from https://collegescorecard.ed.gov/data/.
Over the last 30-40 years, for the poorest wealth group, college attendance has risen. However, overall college graduation has not risen. National graduation average for full-time students at 4-year degree-granting institutions was 59 percent. On a positive note, even though college tuition has risen, students that come from modest backgrounds are attending college at the highest rate ever. Goal: What causes colleges to have higher graduation rates than others? Is it financial background or is there something else causing low graduation rates.
Two models used: Random Forest and Linear Regression
Random Forest:
- Top 5 features explained the vast majority of model performance
- Best hyperparams
- Cross-validated R2 score: 60%
Linear Regression:
- Top features all statistically significant with 95% confidence
- Cross-validated R2 score: 67%
What causes colleges to have higher graduation rates than others?
- Higher educational background of the student's parents.
- Higher tuitions.
- Fewer students transferring into the college.
- Fewer students receiving federal student loans.
- Higher debt for students with high family income.