Vanessa Chang, Jason Ma, Elaina Li, Julia Ganju, Anurag Jakkula, and Li Zhang
The San Francisco Bay Area Chapter held its fifth K–12 student summer program, during which students create and lead projects. The program provides a unique learning experience under the mentorship of chapter officers and volunteers.
Students were assigned mentors and proposed a research idea for their project in June 2023. Over the next three months, mentors met with the students weekly to track the their progress, answer questions, and refine the projects.
The students chose their research projects, obtained the data from publicly available databases, and performed data analysis to test hypotheses using RStudio. At the end of the program, they presented their projects virtually to SFASA members and family.
Elaina Li (10th grader) and Vanessa Chang (12th grader) presented their project on the effect of demographics on homelessness in California before and after the pandemic under the mentorship of Anwen Huang (sophomore at the University of California, Berkeley) and Chenglin Ye (senior principal statistical scientist at Genentech; vice president of SFASA).
They obtained the homeless count data (2017–2022) and downloaded the county population characteristics from the US Census to perform normalization. They found no statistically significant association between the pandemic and homelessness incidences. However, homelessness incidences increased over the study period. Though there was no statistical significance in homelessness incidences between ethnic groups, the older population (age 65+) and males had higher homelessness incidences.
Julia Ganju (12th grader) and Anurag Jakkula (12th grader) investigated the effectiveness of various molecules in inhibiting the aromatase enzyme for breast cancer treatment. They employed machine learning techniques under the guidance of Leah Ung (junior data scientist at the University of California, San Francisco) and Jiying Zou (data scientist at Genentech).
Breast cancer is characterized by a tumor in breast tissue, which is stimulated by the aromatase enzyme’s production of estrogen. Aromatase inhibitors block aromatase activity, reducing the amount of estrogen that promotes cancer cell growth.
The students’ project evaluated 11 regression models to predict aromatase protein inhibition given 252 binary molecular features. Out of the models tested, the LGBMRegressor emerged as the most successful.
They also tested 27 classification models to predict whether the inhibition of the aromatase protein would be above a certain performance threshold. They found LGBMClassifier to be the most effective model in this context.
Jason Ma (10th grader) presented his project on the logistic regression analysis of the team-based game VALORANT under the mentorship of Jeetu Ganju (principal, Ganju Clinical Trials).
Ma aimed to identify which metrics are good predictors of winning for a given player. He found the players who won had higher values of average combat score, kill per round, and assists per round but lower death per round compared to players who lost. These metrics were statistically significant predictors of winning in VALORANT.
At the end of the program, the students had positive comments about their experience and what they learned through participating in it.
“Working on this project with SFASA’s guidance was an amazing opportunity and has inspired us both to continue exploring the field of statistical and machine learning,” said Ganju and Jakkula.
“I learned how to use RStudio to create different types of visualizations of our data and how to use it to run nonparametric tests on our data to see if there is a significant statistical difference between the groups we were comparing,” stated Chang, who had little coding experience.
As someone new to RStudio, Ma learned programming in R and new R packages, including tidyverse, jsonlite, and caret. He discussed how he found it difficult to share his work in a presentation format; however, his mentor gave him a lot of encouragement and suggestions, which helped him complete the project.
Li, who’d returned to the program, stated her previous experience helped her figure out the research questions with appropriate analysis approaches.
SFASA’s summer program not only introduced the students to technical skills but also fueled their passion for statistics. Now a mentor, Huang attended this summer program twice during high school. The hands-on experience, mentorship, and chance to present their work often inspires the students to continue exploring this field.
Leave a Reply