Hospital Pricing App
Unmute
Hospital Pricing App: https://share.streamlit.io/denilsabu/hospital_pricing_app/main/model_integration.py
The dataset:
https://www.dolthub.com/repositories/dolthub/hospital-price-transparency-v2
Data Science:
Our main goal is to create a model that can display/predict the pricing of a medical procedure depending on a users location. The hospitals within the the user's radius search of location will be visualized and there will also be a listing of all the hospitals and their respective prices for the sought after procedure. The hospitals being searched are 23% of all hospitals within the United States and there are over 100,000 unique procedures to search through!
Feel free to connect with us:
Denil Sabu: www.linkedin.com/in/denil-sabu
The dataset:
https://www.dolthub.com/repositories/dolthub/hospital-price-transparency-v2
Data Science:
Our main goal is to create a model that can display/predict the pricing of a medical procedure depending on a users location. The hospitals within the the user's radius search of location will be visualized and there will also be a listing of all the hospitals and their respective prices for the sought after procedure. The hospitals being searched are 23% of all hospitals within the United States and there are over 100,000 unique procedures to search through!
Feel free to connect with us:
Denil Sabu: www.linkedin.com/in/denil-sabu
How much experience does your group have? Does the project use anything (art, music, starter kits) you didn't create?
CodeDay Labs team in the intermediate track.
What challenges did you encounter?
The data set was immense, 165 gb with 100 million rows of data. There was a lot of pruning, converting HCPT to CPT codes, geocoding, and preprocessing in general.
Deployment was also difficult because the dataset was large and Streamlit (the deployment service) had a memory limit. However, after scouring through bugs and trying various services such as git lfs and Amazon Web Services S3, we were able to successfully deploy!
Deployment was also difficult because the dataset was large and Streamlit (the deployment service) had a memory limit. However, after scouring through bugs and trying various services such as git lfs and Amazon Web Services S3, we were able to successfully deploy!
Made with 💖 by robots running our open-source software and GraphQL APIs.