By Caleb Lee and Kimberly Mann Bruch —
Researchers with the University of California San Diego and the Fred Hutch Cancer Center have been working on web-based tools that help individuals visualize how specific coding libraries and data science concepts work in real time. The programs — coined Pandas Tutor and Tidy Data Tutor — illustrate the transformation of data from input to output for programming languages such as Python, R and SQL.
“Data science instructors often spend lots of time pixel-pushing slideshow diagrams to explain dataframe operations to beginners,” said Sam Lau, an assistant teaching professor at UC San Diego School of Computing, Information and Data Sciences’ Halıcıoğlu Data Science Institute. “To help overcome this challenge, we created a set of freely available web-based tools that visualize data table transformations step by step.”
Sean Kross, a staff scientist at the Fred Hutch Cancer Center, said that the team envisioned Pandas Tutor and Tidy Data Tutor as tools to address the difficulty in understanding complex individual code statements and distinguishing between data science tools.
Dense lines of computer code are one way we represent complex data operations, but visual tools give students an alternative and complementary medium for interacting with both code and data,” Kross said.
The web-based tools feature a JavaScript table visualization library and use a language-specific backend that takes related code and adds precise run-time provenance/lineage tracking to it. This way, even relatively complex dataframe operations like pivots and joins can be visualized. Since they were designed for teaching, Pandas Tutor and Tidy Data Tutor support all functions that are typically introduced in introductory data science courses.
“We designed Pandas Tutor and Tidy Data Tutor to show diagrams that illustrate dataframe operations step-by-step by highlighting rows, columns and cells and drawing arrows to show how values were transformed. The diagrams are compact and screenshot-friendly so that instructors can quickly copy and paste diagrams into their slides.” Lau said. “These websites — pandastutor.com and tidydatatutor.com — have been used by instructors at other universities and myself while teaching data science courses. The tools effectively helped students learn code, and student feedback was positive.”
Publicly launched in 2021, pandastutor.com and tidydatatutor.com have attracted over 61,000 users from more than 160 countries. The team has been updating Pandas Tutor and Tidy Data Tutor on a regular basis and has published several conference posters and papers since its original paper, which was titled “Teaching Data Science by Visualizing Data Table Transformations: Pandas Tutor for Python, Tidy Data Tutor for R, and SQL Tutor.”
Lau and Kross’ efforts with the web-based tools are transforming how newcomers engage with data science, and their impact is only just beginning. Lau and Kross continue spreading awareness about these tools to professors worldwide so that they can incorporate Pandas Tutor and Tidy Data Tutor into their classes.
The success of Pandas Tutor has most recently enabled Lau to launch further outreach efforts, including a University of California Historically Black Colleges and Universities (UC-HBCU) grant to host research internships for six students from Morgan State University and Prairie View A&M University at HDSI in Summer 2026.
“We are excited to work with students and professors across the country at Morgan State and PVAMU,” Lau said.
The research was supported by UC San Diego Cognitive Science Professor Philip Guo, the U.S. National Science Foundation (grant nos. IIS-1845900, 1845638, 1740305, 2008295, 2106197 and 2103794) and a grant from Alfred P. Sloan Foundation.
