Case Study: University of Pittsburgh accelerates causal inference research with Google for Education and Google Cloud Techila

University of Pittsburgh runs 1,000 simulations in 48 hours with Google for Education

The University of Pittsburgh, specifically an economics researcher, faced a challenge in running large-scale statistical simulations to assess cause and effect. The on-premises computing resources were too slow, with an initial analysis taking over 30 minutes per simulation, making it impractical to complete the required 1,000 simulations in a timely manner. To overcome this, they turned to Google for Education and its Google Cloud platform, utilizing preemptible instances.

The solution implemented with Google Cloud and its partner Techila Technologies used 10,000 preemptible instances to run the complex simulations in just 48 hours, a task that would have taken over six months on-premises. This approach was not only significantly faster but also more cost-effective, costing less than a third of the on-premises estimate. The results were more accurate, providing the researcher with higher-confidence inferences and transforming the potential for future statistical research across various fields.

University of Pittsburgh