Custom GPTs and R Resources for Statistics Education
Custom GPTs and R Resources for Statistics Education
Purpose: to help other instructors teaching the same course
Common Course ID: MATH 2200, MATH 3200, MATH 3210
CSU Instructor Open Textbook Adoption Portrait
Abstract: This open resource suite is being utilized in a statistics course for undergraduate students by Anjana Yatawara at California State University, Bakersfield. The resources include a custom GPT assistant for learning R (MATH 2200), a practice-focused AI tutor using course notes and problems (MATH 3200), and a Shiny App Builder assistant (MATH 3210). The main motivation to develop custom GPTs was to increase accessibility, support student learning with tailored tools, and eliminate textbook costs. All of our students access the GPTs online through the ChatGPTEDU platform and course LMS.
Course Title and Number:
MATH 2200 - Introduction to Statistics
MATH 3200 - Mathematical Statistics
MATH 3210 - Statistical Computing and Graphics
Brief Description of course highlights:
- MATH 2200: Introductory statistics focusing on data analysis, hypothesis testing, linear regression, and experimental design. Includes R coding basics.
- MATH 3200: A rigorous exploration of probability theory and statistical inference.
- MATH 3210: Hands-on course focused on data visualization, reproducible workflows, and building R Shiny apps.
Student population: Primarily undergraduate students in statistics, data science, or related majors. MATH 2200 is taken by many students fulfilling general education requirements, often with little to no prior coding experience. MATH 3200 and 3210 serve statistics majors with background in calculus and probability.
Learning or student outcomes:
- Interpret and communicate statistical findings
- Use R for data analysis and modeling
- Apply theoretical concepts in statistical inference
- Create interactive data tools using Shiny (MATH 3210)
Key challenges faced and how resolved: Many students struggle with R programming and interpreting statistical outputs. The custom GPTs were developed to provide instant, interactive support tailored to each course’s content. They help students debug code, reinforce concepts, and engage with materials more actively.
Instructor Name - Anjana Yatawara
Assistant Professor in Statistics at CSU Bakersfield, Department of Mathematics
Please provide a link to your university page.
https://www.csub.edu/math/facstaff/faculty.shtml
Please describe the courses you teach
- MATH 2200: Introductory statistics focusing on data analysis, hypothesis testing, linear regression, and experimental design. Includes R coding basics.
- MATH 3200: A rigorous exploration of probability theory and statistical inference.
- MATH 3210: Hands-on course focused on data visualization, reproducible workflows, and building R Shiny apps.
Describe your teaching philosophy and any research interests related to your discipline or teaching. I focus on making statistics accessible and engaging through interactive tools and open resources. My research explores how AI, particularly large language models, can be used as structured tutors in statistics and data science education.
OER/Low-Cost Adoption Process
Provide an explanation or what motivated you to use this textbook or OER/Low Cost option. Support learning beyond the classroom, save students money, offer scalable personalized support, integrate AI tools into statistical education in a structured way.
How did you find and select the open textbook for this course? Developed by the instructor using personal teaching materials and feedback from students. Also reviewed AI best practices and consulted with colleagues.
Sharing Best Practices:
- Start small with one topic or assignment
- Will provide guide on how to use these tools where necessary
- Provide students with a usage guide to ensure ethical and productive AI use
- Customize GPTs to match your course content for better alignment
- Engage students in using AI thoughtfully, including debugging, think-aloud protocols
Describe any key challenges you experienced, how they were resolved and lessons learned. Keeping GPTs updated as course content evolves, teaching students how to use AI responsibly, and ensuring accessibility and inclusivity in AI tools.
Custom GPT Assistants for Statistics Courses
Brief Description:
- MATH 2200: A GPT-based assistant helps students learn R syntax, interpret errors, and apply statistical methods using the `mosaic` and `BSDA` packages.
- MATH 3200: A practice-oriented GPT built from class notes, sample problems, and conceptual reviews.
- MATH 3210: A GPT designed to assist students in building R Shiny apps, offering examples, code templates, and debugging help.
Please provide a link to the resource
https://chatgpt.com/g/g-67ef41eed5c48191bae80bf9fe6670a1-promptperfect-prompt- engineering-gpt-dr-y
https://chatgpt.com/g/g-67e96d4ea070819197453379407cb057-r-learning-assitant-gpt- dr-y
https://chatgpt.com/g/g-67e96f613f7481918e13c4764c2d4348-python-learning- assistant-gpt-dr-y
Authors: Anjana Yatawara
Student access: Students access GPTs through shared links and LMS announcements. GPTs are hosted on the ChatGPT platform and require no download/purchase.
Supplemental resources:
- Sample R scripts, lecture notes, recorded walkthroughs
- Custom prompts to support problem-solving and review
- Faculty-only customization features in GPTs
Provide the cost savings from that of a traditional textbook. No cost savings. Out MATH 2200 is already taught using open source resources.
License: The GPTs are shared openly via the ChatGPTedu platform and built using course materials created by the instructor.