# Portraits of Graduate Advising: Advisor Rating Correlates and an Evidence-Based Rubric for PhD Applicants
> Analysis of 11,311 advisor profiles and 12,344 bilingual reviews from the OpenAdvisor platform. Ratings are strongly bimodal; professional metrics (h-index, citations) explain essentially no rating variance; keyword theme classification over-identifies themes in Chinese text by ~35x versus LLM classification. Produces a 10-dimension, behavior-only advisor-selection rubric and an interactive advisor helper tool that outputs a preliminary read plus a structured prompt for LLM-assisted analysis.
Type: paper
Venue: Preprint
Date: 2026
Authors: Dixi Yao
## Direct Links
- [Canonical topic page](https://dixiyao.github.io/topics/cs-ml-advisor-helper/)
- [Project page](https://dixiyao.github.io/csmladvisor/)
- [Paper](https://dixiyao.github.io/assests/papers/portraits-of-graduate-advising.pdf)
- [PDF](https://dixiyao.github.io/assests/papers/portraits-of-graduate-advising.pdf)
- [Code](https://github.com/dixiyao/CS-ML-Advisor-Helper)
- [Markdown alternate](https://dixiyao.github.io/csmladvisor/index.html.md)
## Search Queries and Aliases
- CS-ML-Advisor-Helper
- CS ML advisor helper
- advisor helper
- PhD advisor selection tool
- advisor selection rubric
- portraits of graduate advising
- OpenAdvisor analysis
- choose PhD advisor
- advisor red flags
- graduate advising rubric
## Search Intents
- how to choose a CS or ML PhD advisor
- evidence-based advisor selection rubric
- are advisor review platforms reliable
- PhD advisor red flags checklist
- interactive tool to evaluate PhD advisor
- LLM prompt to analyze a potential advisor