How this page was made: It was generated from real, de-identified teaching evaluations collected over many years. The commentary, explanatory text, and visuals were generated with GPT-5.6 Sol High during roughly 30–60 minutes of discussion with Ben, much of it spent refining small design details, hover information, clarifying language, and explanations like this one.
Make the complex feel possible.
Across analytics, visualization, machine learning, and organizational behavior, one pattern holds: students describe rigorous material becoming practical, clear, and human.
This was truly the best class I have had at Tepper. No other class has been as good as this.Anonymous student, Data Visualization · Fall 2025
The good trends
These are not isolated high points. The strongest recent stories persist across sections and grow over time. Select a story to isolate it in the timeline.
Every section, in context
Each dot is one section. Dot size reflects responses; the line is the response-weighted yearly average. Career chapters appear as background shading so the teaching record stays in the foreground.
Explore the dots:Hover over any dot for course, term, section, rating, and student-comment details. Click or tap a dot to keep its full details below the chart.
The teaching signature
Recent results are strongest where the comments are strongest: respect, relevance, explanation, and care.
Respect is the clearest through-line.
Across 509 responses since 2025, “shows respect for all students” is the highest-rated dimension. Comments repeatedly add patience, approachability, and willingness to revisit difficult material.
Constructive ideas to keep improving
Students’ most useful suggestions are specific and actionable. These exact comments were selected because they explain what could change while remaining respectful and solution-oriented.
Mention counts come from the keyword-assisted coding pass; comments may contribute to more than one area. Selected quotations are illustrative, not exhaustive.
Courses, with student voices
Ratings give the contour; the quotes show what the experience felt like. Each quote is anonymous and drawn directly from the supplied evaluations.
Informal survey feedback from select CMU-Q executive-education courses
Only two locally retained 2014 surveys were available. These courses used an informal survey outside CMU’s formal Faculty Course Evaluation system, so this is not a complete record of all executive-education classes taught. The programs no longer run, and the current location of any missing historical records is unknown. Email archives were not available for verification.
Private by design
This page reports only Benjamin Collier’s ratings and anonymous student comments. It includes no colleague ratings, institutional comparison averages, student names, or student email addresses.
- Comparable five-point sections are weighted by the number of responses to each item.
- Duplicate PDFs were de-duplicated by file hash.
- Reports using a different ten-point instrument are shown separately and never mixed into the main averages.
- No files were present for 2017–2022; the gap is left visible.
AI-assisted analysis and authorship
This report was prepared with GPT-5.6 Sol (High reasoning) in the Academic edition with FERPA-enabled features. The model worked from de-identified course-evaluation data.
Supported comment extraction and normalization, keyword-assisted thematic coding, grouping into strengths, neutral observations, and improvement areas, candidate quote selection, and drafting the report narrative and page language.
Ratings, response counts, response-weighted averages, and time trends were calculated directly from the source records. The model did not generate or estimate missing ratings.
Displayed student quotations retain the supplied source wording rather than being rewritten by AI. Student identifiers were excluded, and colleague names and individual colleague ratings were not included in the report dataset.
Public verification note
The displayed totals and response-weighted averages were recalculated from de-identified section results before publication. The public page intentionally excludes institutional comparison fields and private source-document hashes. The private audit remains available to the author.
What students notice
A keyword-assisted coding pass across 370 anonymous comments surfaces the recurring ideas. A comment may contribute to more than one theme.
What already lands
Engagement (105 mentions), applied learning (99), and care/accessibility (86) appear most often. Students value a practical classroom with visible enthusiasm and human support.
What to keep tuning
Pace/workload (49 mentions), organization (38), and feedback/grading (29) recur as improvement themes, especially in compressed technical courses. These are mention counts, not satisfaction scores.