The ASU Library Unit for Data Science hosted the Data Science Student Open Project Showcase on Friday, bringing students, faculty and community members together to explore how data can drive positive change.
At Hayden Library’s third-floor Research Commons, the event highlighted new opportunities at the intersection of research, technology and community impact. Throughout the showcase, students from the Open Projects program presented their findings and solutions surrounding real-world problems.
At the showcase, Yilin Che, a third-year PhD student studying political science, presented a data science project using a machine learning algorithm as a method to help predict chronic kidney disease.
Drawing data from the UC Irvine Machine Learning Repository, Che used many machine learning algorithms such as logistic regression, Naïve Bayes, decision tree and random forest to identify key risk factors like diabetes, high blood pressure and family history for kidney disease prediction.
"Usually, using the manual was a time-consuming process, but by using machine learning, and mainly the supervised or unsupervised learning approaches, we can really accelerate the treatment and the testing of CKD in conjunction with the traditional methods," Che said. "Machine learning can replace the traditional method, but it is really a complement."
Shreyas Ramani, a graduate student studying data science, analytics and engineering, took a slightly different approach to the CKD data. Ramani divided a data set into about 75% training data, then used a supervised learning machine to predict the rest of the set.
"I'm trying to see if any of the variables are really doing a good job to predict whether someone's going to have a kidney failure," Ramani said.
Another project at the showcase, led by Likhitha Geddam, a graduate student studying data science, analytics and engineering, analyzed another UC Irvine dataset to understand patterns among critical care patients with diabetes. Using data on diabetes from the UC Irvine repository, the team focused on identifying clusters and outliers to reveal relationships between varying features.
The team worked through challenges with the data to organize clusters of patients with similar characteristics. Geddam said they had to use different processing methods to take into consideration null values. However, they were ultimately successful in separating characteristics using the machine learning model.
Projects at the event used predictive modeling to forecast hospital readmissions among patients with Type 2 diabetes.
In addition to student presentations, attendees visited tables from ASU Library departments that showcased data visualization tools, open-access resources and research services. Student organizations such as Women in Data Science and Statistics at ASU were also present, encouraging collaboration and networking among attendees.
"Machine learning algorithm is a really powerful tool for both the data science or social sciences ... to assess the decision-making process," Che said.
Edited by Kate Gore, George Headley and Ellis Preston.
Reach the reporter at afrahma1@asu.edu.
Like The State Press on Facebook and follow @statepress on X.
Ariana is a sophomore studying Biomedical Informatics. This is her first semester with The State Press.
