Researched the application of natural language text processing in replacing the patient-based surveys that Houston Methodist Hospital administers.
This research project aimed to solve the issue that is present with current methods of obtaining data on patient satisfaction. Often times, patients who were given the surveys tended to spend very little time providing feedback due to the fact that the forms were very lengthy. Some would even forget about the forms and would remember them 6 months later, delaying the time it takes for the information to reach the health care institution.
Knowing about the massive amounts of criticism that are relayed over review websites, such as Yelp and YellowPages, we set out to retrieve these unstructured data from the web using Scrapy, a data extracting python framework, and to conduct sentiment analysis on said data. This allowed us to provide real time data to the hospitals, allowing them to make informed decisions on how to improve their overall healthcare.
By using a web crawler and conducting sentiment analysis on the retrieved data, we were able to provide a better alternative to the patient-based surveys, providing Houston Methodist real-time data on patient satisfaction.
There are many other applications of this research, such as extending this tool to fit more than just the domain of health care institutions. This research can also be used to create tools for companies that are looking to get better feedback on how they can improve their services.