Reproducible Research Project for Institution Retention Rates

Binder link:

Github page: https://ucb-stat-159-s23.github.io/project-Group11/

Dataset link:

Project Goal

This is a project attempting to analyze the factors that affect student retention at universities in a reproducible manner. We conduct exploratory data analysis, compute feature importance, and implement logistic regression for prediction. The data originates from the U.S. Department of Education College Scoreboard through the Institution-level data files for 1996-97 through 2020-21 containing aggregate data for each institution. The dataset includes information on institutional characteristics, enrollment, student aid, costs, and student outcomes. You can find more information here: https://collegescorecard.ed.gov/data/

Useful Commands

• To run all tests for utility functions, run pytest tools from the terminal in the root directory.
• To create the environment, run make env from the terminal in the root directory.
• In addition, you may run make html to build the jupyter-book html, make clean to remove figures and html files, and make all to run all the aforementioned make commands.

Repository Structure

• /data: cleaned data csv files
• /figures: all generated figures as png files
• Jupyter Notebooks:
  • main.ipynb: the main narrative notebook of the research project
  • EDA.ipynb: introduction, data description and some basic exploratory data analysis of retention rates
  • EDA_control_of_school: the code of the control of the school for exploratory data analysis and EDA related figures
  • EDA_in_out_state_tuition: the code of in/out state tuition & fee for exploratory data analysis and for EDA related figures
  • EDA_rece_loans: the code of individuals receiving federal loans for exploratory data analysis and for EDA related figures of percentage
  • EDA_parent_edu: the code of parent education for exploratory data analysis and for EDA related figures
  • EDA_program_offered: the code of programs offered for exploratory data analysis and for EDA related figures
  • EDA_num_var: feature analysis on numberical variables
  • Variable_Analysis_1.ipynb: code for logistic regression model and their corresponding figures of four-year institutions retention rates
  • Variable_Analysis_2.ipynb: code for logistic regression model and their corresponding figures of less than four-year institutions retention rates
• Python utility package tools:
  • /tools: code and tests for python package
  • Setup files: setup.py, setup.cfg, pyproj.toml
• Environment files: environment.yml, envsetup.sh
• Jupyter Book: _config.yml, _toc.yml, conf.py, postBuild, requirements.txt
• contribution_statement.md: authors' contributions

Dataset Attribute Information

• UNITID: Unit ID for institution
• CONTROL: Control of institution
• CCUGPROF: Carnegie Classification -- undergraduate profile
• CCSIZSET: Carnegie Classification -- size and setting
• ADM_RATE: Admission Rate
• SAT_AVG: Average SAT equivalent score of students admitted
• UG: Enrollment of all undergraduate students
• UGDS_[...]: Total share of enrollment of undergraduate degree-seeking students who are [...]
  • UGDS_WHITE (white), UGDS_BLACK (black), UGDS_HISP (Hispanic), UGDS_ASIAN (Asian), UGDS_AIAN (American Indian/Alaska Native), UGDS_NHPI (Native Hawaiian/Pacific Islander), UGDS_2MOR (two or more races), UGDS_NRA (non-resident aliens), UGDS_UNKN (unknown), UGDS_WHITENH (white non-Hispanic)
• NPT4_PUB: Average net price for Title IV institutions (public institutions)
• NPT4_PRIV: Average net price for Title IV institutions (private for-profit and nonprofit institutions)
• TUITIONFEE_IN: In-state tuition and fees
• TUITIONFEE_OUT: Out-of-state tuition and fees
• AVGFACSAL: Average faculty salary
• PCTPELL: Percentage of undergraduates who receive a Pell grant
• RET_FT4: First-time, full-time student retention rate at four-year institutions
• RET_FTL4: First-time, full-time student retention rate at less-than-four-year institutions
• PCTFLOAN: Percent of all undergraduate students receiving a federal student loan
• PAR_ED_PCT_MS: Percent of students whose parents' highest educational level is middle school
• PAR_ED_PCT_HS: Percent of students whose parents' highest educational level is high school
• PAR_ED_PCT_PS: Percent of students whose parents' highest educational level was is some form of postsecondary education
• DEP_INC_AVG: Average family income of dependent students in real 2015 dollars
• IND_INC_AVG: Average family income of independent students in real 2015 dollars
• GRAD_DEBT_MDN: The median debt for students who have completed
• WDRAW_DEBT_MDN: The median debt for students who have not completed
• FAMINC: Average family income
• MD_FAMINC: Median family income
• PRGMOFR: Number of programs offered

For more information on the additional variables, please refer to the Data Dictionary. Make sure to download the Most Recent Institution-Level Data data dictionary and you can find all the descriptions of the variables in the Institution_Data_Dictionary tab.

Testing

1. To guarantee that the correct environment is in place, execute the make env command. If an outdated version of the environment is already set up, or if you need to modify the environment, activate a separate environment using conda activate and then execute make remove-env first to remove the prior setup.

2. After running conda activate final_proj, run conda install -c anaconda pytest.

3. Run pytest tools.

Running Instructions

To run the main notebook, you have two options: you can either use the provided binder link, or you can clone the repository and run the notebook locally. If you choose to run the notebook locally, we have also included a makefile with instructions to help streamline the process.

License

The project is released under the BSD 3-clause License.