AWS Business Intelligence Blog
Amazon QuickSight partners with Stanford’s SHTEM internship to empower students in generative AI and research
Founded in 2019, Stanford University’s SHTEM (science, humanities, technology, engineering, and mathematics) Program, created by the Stanford Compression Forum, is a summer program intended to provide high schoolers and community college students from diverse backgrounds early exposure to collaborate and conduct research that transcends traditional disciplinary boundaries. This year, students conducted research on topics ranging from the detection of pulsars to the reduction of power loss in electrical circuits. One of the larger goals of the program is to break down barriers to entry for underrepresented students into SHTEM-related fields.

Presentation Day from Summer 2024
This summer, Amazon QuickSight partnered with Tsachy (Itschak) Weissman, a Stanford Engineering Professor and one of the founders of the SHTEM Program, to integrate generative AI and data visualization learnings into the program by introducing students to Amazon QuickSight and Amazon Q in QuickSight. The goal was twofold: to equip the student researchers with valuable data analysis and visualization skills by the end of the summer and ignite their enthusiasm about the possibilities of generative AI.
In this post, hear from four students and mentors about their experience using Amazon QuickSight this summer during Stanford’s SHTEM internship.
David’s experience
David Jose Florez Rodriguez, one of the SHTEM Program’s mentors and coordinators, spoke about his experience leading and mentoring his team while also learning about the power of Amazon Q in QuickSight. His team chose to work on a computer vision project focused on detecting and segmenting guns from images as a response to the nation’s issue with gun violence. By putting their training data into Amazon QuickSight, his students were able to gain insights from their data, which equipped them to determine which family of models were best for their project.
David decided to join SHTEM as a mentor and coordinator because he found that the program “finds students who are extra driven and helps bridge that gap from what the public school system offers to what [they are] going to need to succeed in university.”
David expressed his appreciation for the QuickSight beginner-friendly features, including its low-code, no-code option for data analysis, which freed up time for his team’s core research. According to David, “I would have had to teach them about plotting libraries and a little bit of data analysis and debugging,” but with QuickSight, “we could just skip over all that, which was really nice.” He believes that, with QuickSight, the students without technical expertise were able to quickly focus on their project because the tool provided them accessible data analysis capabilities.
Khushi’s perspective
Khushi Upadhyay is a high school senior who participated in the SHTEM Program this summer. Her team’s research project focused on pulsars, which are the remnants of dying or exploding stars that emit beams of electromagnetic waves. They had a few goals for their research:
- Analyze the data collected by the Green Bank Telescope to try and discover new pulsars,
- Make pulsar data more accessible and comprehensible to nonexperts to increase overall awareness of pulsars
- Develop a machine learning (ML) algorithm to classify plots as representing pulsars or not
- Analyze data on previously discovered pulsars to look for any new patterns or trends that weren’t evident before, using data of all currently known pulsars from the Australia Telescope National Facility (ATNF) Catalog.
Initially, Khushi was unsure if her team would be able to use QuickSight to effectively analyze scientific data because she had previously seen it used in business and sales analytics. However, she was “ecstatic to find out how seamlessly compatible QuickSight was with the data and how broad its use cases could be.”
Khushi’s creative spirit shone through when she elaborated on her experiences learning to use QuickSight. She explained how the learning process itself was almost equally as fun as making the dashboards because she was able to explore the tool and learn by doing. In addition to taking the QuickSight author courses (part 1 and part 2), she also used helpful tools, such as DemoCentral, where users can explore features of QuickSight through sample dashboards, and the QuickSight Community, an online resource where QuickSight users ask and answer questions, access learning resources, and find out about events. Khushi also got inspiration from the QuickSight Arena Gallery, explaining that “by looking at what other people are doing and what kind of functionalities they’ve built into their dashboards, I was able to get a lot of ideas about what to do with mine.”

Pulsar Feature Analysis QuickSight Dashboard
Khushi’s team was able to conduct an in-depth data analysis of all approximately 3,000 currently known pulsars, encompassing over 60 astrometric parameters, and more than 15 additional calculated fields. This analysis was extremely successful in that her team managed to construct visuals and plots of over 20 major pulsar features against each other, identifying meaningful trends, patterns, and critical anomalies. The ability to refine visuals with Amazon Q in QuickSight was a game changer in creating detailed representations of complex scientific data while forging beautiful dashboards. A crucial example was the ability to plot logarithmic graphs, which are vital in displaying certain relationships, and vary the axis step count.

Pulsar Surface Magnetic Field QuickSight Dashboard
Her team was able to use Amazon Q in QuickSight heavily in their research project this summer. She was impressed by the generative AI capabilities of Amazon Q in QuickSight, specifically by the ability to use natural language to ask questions about her data and identify insights in minutes. She said that QuickSight “exceeded [her] expectations both in terms of usability and functionality in the sense that it offers a robust and intuitive solution to really harness the power of data.”

Pulsar Age QuickSight Dashboard
By using QuickSight, Khushi and her team were able to accelerate their progress toward their two main learning goals. First, the use of dashboarding equipped her team to dive deep into their understanding and gaining a comprehensive view of pulsars and their metrics. Second, using the data storytelling capabilities of QuickSight, her team engaged users and provided a seamless user-data interaction, fulfilling their goal of creating accessibility and comprehensibility for pulsar data and insights. For their team, it was a key priority to give users more ways to discover and explore pulsar data while keeping the user experience straightforward and intuitive.
Ellie’s embrace of QuickSight
Ellie Han, a high school senior in the SHTEM Program, worked with Khushi analyzing data to discover and classify pulsars, using QuickSight to uncover key insights about their findings. “QuickSight did a really good job of helping us visualize the data,” Ellie explains.
Through interactive dashboards and charts, her team was able to visualize and analyze pulsar data more effectively. Ellie’s team created scatter plots to compare the dispersion measure (DM)—the amount of gas between Earth and the pulsar—with the frequency and location of pulsars. By examining these relationships, Ellie and her team could estimate how far away a pulsar was and determine if a signal was likely to be from a real pulsar. Higher DM values suggested more distant pulsars, and comparing frequency and DM helped distinguish genuine pulsar signals from other space phenomena. This visualization approach allowed Ellie to quickly understand complex pulsar data, improving her ability to identify these unique celestial objects.
Although Ellie faced some initial challenges in navigating the QuickSight features, she was able to get up to speed with the help of her project mentor and online resources, such as the QuickSight author courses, and browsing the Amazon QuickSight website.
Though time constraints prevented her from fully exploring AI and ML capabilities of QuickSight, Ellie was impressed by the platform’s potential to accelerate data-driven discovery. “I didn’t know how involved AI was with research in general,” she says. “It was very mind-blowing to me.”
As a next step, Ellie is independently conducting research that involves detecting gravitational waves through pulsars, taking the research a step further from finding pulsars to using them to find gravitational waves. She is also interested in taking more Amazon Web Services (AWS) courses and gaining certifications. Overall, Ellie’s experience demonstrates how powerful data visualization and analytics tools can be in empowering students to tackle complex research challenges.
Manya’s insights
Manya Singla applied for the SHTEM internship because she was looking to gain experience in research. With her team last summer, Manya aimed to determine if incorporating wide-bandgap semiconductors into electrical circuits could reduce power loss. Wide-bandgap materials are known to maintain efficiency even in high voltage or high temperature environments. Manya and her team decided to use QuickSight to help tell a story about their research data.
Because she was new to Amazon QuickSight, Manya took time to experiment with the many different features, including Amazon Q in QuickSight. She found that playing around with the data in QuickSight helped her understand the research her team was doing on a deeper level. For example, she explained that QuickSight made it incredibly simple to see the effect of consistent temperatures on different attributes of the circuits by switching the variables of speed and power loss. When it came time to write up her final report, Manya’s experiences with QuickSight equipped her to approach the analysis with a lot more insight.
Another aspect of QuickSight that Manya found important was its ability to handle the large datasets they were working with through the Super-fast, Parallel, In-memory Calculation Engine (SPICE), which is the robust in-memory engine that Amazon QuickSight uses. It’s engineered to rapidly perform advanced calculations and serve data. Using SPICE, Manya’s team could focus on experimenting with their data faster.

Manya’s using QuickSight visuals to experiment with relationships between research variables
Manya found SHTEM to be a very valuable learning experience. As she explains, “when it comes to end products, there can be a bit of pressure to ensure that everything is perfect and that the final result is the highlight and focus of your research, because that’s what people are reading it for,” but when it comes to internships like SHTEM, “their educational value is really great because your professors and mentors will encourage you to showcase your learning process in the actual output you are producing.” Using Amazon QuickSight, Manya learned about the importance of clarity and simplicity in data visualization. She found that being able to “showcase data in a meaningful way” was crucial because it “increased the reach of [her] research.” Manya recognized that effective data visualization played a key role in communicating her findings with impact.
Empowering the next generation of researchers
The partnership between Amazon QuickSight and Stanford’s SHTEM Program has empowered a diverse group of student researchers to explore the power of data visualization, analytics, and generative AI. Professor Weissman emphasized the program’s objectives and outcomes:
“By integrating Amazon QuickSight into the SHTEM program, we aimed to empower students with cutting-edge data visualization and generative AI tools. The intuitive QuickSight features allowed them to uncover meaningful insights in their projects and communicate their findings effectively, all while gaining working knowledge and appreciation for the transformative potential of these technologies.”
The program has not only equipped participants with state-of-the-art tools, but has also ignited their enthusiasm for the potential of these technologies to revolutionize research across disciplines. By providing hands-on experience with QuickSight, the internship has helped foster the next generation of data-driven problem solvers and innovators, preparing them to tackle complex challenges in their future academic and professional endeavors.
Acknowledgements
Sincere gratitude to the Stanford SHTEM internship team, particularly Professor Tsachy Weissman and program coordinators David Jose Florez Rodriguez, Sylvia Chin, Suzanne Marie Sims, and Lucrecia Kim-Boswell. Their efforts ensured the seamless integration of Amazon QuickSight into the program. Special thanks to David Jose Florez Rodriguez, Khushi Upadhyay, Ellie Han, and Manya Singla for sharing valuable insights on QuickSight usage. Lastly, thank you to Jill Florant for initiating this collaboration with Stanford and her instrumental role in its implementation.
About the Author
Julia Bernstein is a Program Manager for Amazon QuickSight, a cloud-centered, fully managed BI service, and the Developer Liaison for the Amazon QuickSight Community.