MALVERN, Pa. — Ruchika Chari and Malavika Mathur, two full-time Master of Information Science students at Penn State Great Valley, are preparing to graduate. While this can be a stressful time for many, both women have secured job offers, thanks to their studies combined with hands-on research experience with Satish Mahadevan Srinivasan, assistant professor of information science.
Chari, an international student from India, has an undergraduate degree in computer science. Chari decided to attend Penn State Great Valley because the information science curriculum closely aligned with her interests and career aspirations. Mathur, from Thailand, received a bachelor’s degree in telecommunications from University Park. After working as an iOS developer in the Philadelphia area for a few years, Mathur decided to enroll full-time at Great Valley.
Looking to gain experience outside the classroom, Chari asked Srinivasan if there were any research opportunities available. Srinivasan brought this up with the chancellor, and was able to offer Chari a part-time research assistant position in the new big data lab.
This fall, Srinivasan formed a five-person team to work in the big data lab: Chari; Mathur; Aureo Zanon, a World Campus student in data analytics; Tianhai Zu, a master of finance graduate; and another master of information science student, Priyangana Sharma.
Over the course of six weeks, the team captured and stored 24 million tweets across 15 states — equally made up of Democratic, Republican, and swing states — to understand the public’s emotions toward Hillary Clinton and Donald Trump. The team used several machine learning techniques to classify the tweets: first into four main emotion types — sadness, happiness, trust and disgust, and then into positive or negative sentiments. They used Tableau, a data visualization tool, to look at the results.
Prior to the election, the team noticed fluctuations in the public’s feeling toward the candidates. After a video surfaced in which Trump made vulgar comments about women, many of the public’s tweets showed disgust. But after the third debate, the data revealed Trump was gaining trust.
In an interesting turn of events, on Nov. 6 Obama said if Clinton won the state of Florida, she would win the election. But based on tweets, Srinivasan and his team predicted Trump would win Florida — and he did. Out of the 15 states monitored, the team predicted the outcome of eight correctly including the swing directions for Florida and Ohio.
“I was fortunate to work with such a strong and motivated team. While this was a time-consuming project with a huge volume of data, they had great communication skills, making it easy to execute tasks,” said Srinivasan. “Each person added something different. Chari has a background in computer science, while Mathur has relevant work experience.”
Zanon and Zu, both planning to pursue doctoral degrees in the near future, assisted with data collection and cleaning. Sharma provided overall support to the group.
In addition to Srinivasan’s team, other Great Valley students participate in research with faculty members. The collective group frequently meets to discuss each other’s work.
“We really got to know the other projects,” said Chari. “Even though we are researching different things, we can really learn a great deal from one another — the different tools, ways of analyzing data.”
“It became a collaborative environment,” agreed Mathur. “We sought feedback and ideas from each other, which mimics how corporate teams work together in the real world.”
The comradery of the big data lab is also important to Srinivasan. “I want to create a close-knit group where students can talk about their research and academic work,” he said. “As faculty, we benefit from the lab community. Everyone contributes, so I am able to learn from my students, too.”
Chari, who graduates this December, will move to Pittsburgh to work as an associate consultant at Highmark. Once Mathur graduates in May, she will relocate to Long Island for a job with PASSUR Aerospace to monitor aviation delays and traffic flow.
Grateful for their experience at Great Valley, both women feel well-prepared for their jobs.
“Our courses built a strong foundation, working hand-in-hand with our time in the big data lab,” said Mathur. “I learned how to manage large amounts of data, and I feel confident to take on challenges in my new role.”
“My experience using Tableau gave me a huge advantage in my job search,” said Chari. “It allowed me to look at data in a new way and see the bigger picture.”
Srinivasan plans to publish the research on emotion classification, visualization and analysis with fellow faculty member Raghu Sangwan. By the next election, he hopes sentiment analysis will be more automated and provide a real-time snapshot of the public’s emotions.
For students seeking beyond the classroom, Srinivasan encourages them to talk to their professors. “Don’t be afraid to approach faculty to ask for research opportunities. We have a lot of ideas, and are often looking to collaborate with students.”