Given a data set with 500 samples and the following eight variables:
GRE (GRE scores), TF (TOEFL scores), UR (university rating), SOP (Statement of Purpose strength), LOR (Letter of Recommendation strength), GPA (undergraduate GPA), Research (have research experience or not), Admit (chance of admittance),
we are interested in the causal relationships between them, especially the causal influence on the variable Admit (the chance of admittance).
The main methodology in this project is the use of causal graphical model which is a Bayesian network where the directed edges indicate the causal relationships between nodes.
I established the model by making reasonable causal assumptions based on our prior knowledge and intuition about the variables.
/graduate_admissions contains two datasets of graduate school admission dicisions.
/CausalAdmit_Joy.ipynb presents my work on how I constructed a Bayesian network, then explored causal relationships between different variables.