Jacqueline Krohn

This project was a fun way for me to play with some APIs and unstructured data and then try out some sentiment analysis with finbert, a pre-trained NLP model to analyze sentiment of financial text.

The Notre Dame Capstone team was tasked with improving the patient experience for the different divisions and facilities that make up Children’s Mercy Kansas City. Over the course of four months, our team was able to identify several “time-sensitive divisions” (health divisions where patients were most affected by wait time and scheduling time). This sensitivity, coupled with longer-than-desired wait and scheduling times, has led to a low patient experience (indicated by low facility ratings) for various Children’s Mercy divisions. Given this, our team was determined to answer—“What can Children’s Mercy do in order to improve low-performing divisions?” After running many different models, our findings indicated that just a 5-minute change in wait time at those sensitive divisions improved positive ratings by nearly 5% while reducing scheduling times at those sensitive locations increased positive facility ratings up to 16.8%.

In order to reduce the potential for power distribution line related disasters, sensing equipment can be attached to the distribution line. These sensors can read voltage and current in the line, and by transmitting this data back to a central location, an updated state of the power distribution system can be easily monitored. By being able to monitor locations of damaged or downed lines, quicker response and maintenance can be completed resulting in a safer power grid.