Deepti Saravanan and Jahnavi Swetha Pothineni (Center Left and Center Right) pose with Courant Director Russel Caflisch (Left) and Professor Anasse Bari (Right) at the Prizes & Fellowship Ceremony.

Alumni Spotlight: Deepti Saravanan and Jahnavi Swetha Pothineni Use Predictive Analytics to Track Temperature Rise

October 24, 2023

By Sarah Ward

When it comes to measuring climate change, much of the existing research focuses on the effect of higher temperatures on the natural world—tracking rises in sea level, ice melt, and carbon emissions for example. But a recent study published by the Institute of Electrical and Electronics Engineers focuses instead on tracking the human factors at play. The paper’s co-authors, Deepti Saravanan and Jahnavi Swetha Pothineni, were second year students in Courant’s Masters in Computer Science program when they were inspired to begin the project as an assignment in Professor Anasse Bari’s Predictive Analytics course. 

Deepti and Jahnavi decided to take a “different analytics approach using new alternative data sources” by tracking human activities—including internet usage, meat consumption, air travel, and car sales—to forecast average surface temperatures across regions and decades. They trained four predictive models on the datasets which indicated varying levels of correlation between human behavior and temperature rise. Deepti and Jahnavi describe the data collection as “a prominent hurdle” in their research process. Professor Bari had warned that 75% of time spent on any AI project is devoted to data preparation. By drawing upon various publicly accessible sources, Deepti and Jahnavi were eventually able to collect a “meticulously curated dataset that holds the potential to streamline efforts of researchers and data scientists engaged in analogous pursuits.” 

The results indicated that the human behavior most closely linked to temperature rise is carbon consumption followed—less expectedly—by fertilizer consumption. The authors are hopeful that this research will lead to more informed decisions on climate action. “We hope our study will encourage further analysis focusing on human activities,” Deepti and Jahnavi said, “this is imperative as these are controllable actions which could go a long way in combating climate change.”

Applying Predictive Analytics to Climate Change: Predicting Temperature Rise Using Human Behavior Alternative Data” was accepted and presented at the Swiss Conference on Data Science in Zurich this past summer. “This is a milestone in our career as women in tech,” the authors said, “We are motivated more than ever to pursue research, especially discovering and presenting insights on topics that matter.” Deepti and Jahnavi were also honored with the Suzanne McIntosh Fellowship at Courant’s Prizes & Fellowships Ceremony in May. The late Professor McIntosh was a leader in AI research and an advocate for women in science; Deepti and Jahnavi “feel honored to carry her name and her legacy.”

“Courant instilled a sense of belonging in us,” Deepti and Jahnavi recall. “We have gone from ideas, to experiments, to deployment of predictive analytics, to publishing our work at a leading conference.” The two will carry this knowledge and confidence into their newly established careers. After graduation in the spring, Deepti accepted a position as a Data Scientist at Aetna, a CVS Health Company; Jahnavi currently works as a Software Engineer at Bloomberg.

Photo: Deepti Saravanan and Jahnavi Swetha Pothineni (Center Left and Center Right) pose with Courant Director Russel Caflisch (Left) and Professor Anasse Bari (Right) at the Courant Prizes & Fellowship Ceremony in May 2023, courtesy of Anasse Bari.