Hospital beds finder

A full stack solution that maps a USA location of your choosing with hospitals colorcoded to display their percentage of available beds. It uses the USA Definitive Healthcare | Hospital Beds datasets.

Comment

hospital-beds

USA hospital beds capacity web app. Try it at http://hospitalbeds.link/

Inspiration

The COVID-19 lock down has us thinking twice before going out and thinking more than a couple times before going to a hospital if not in an emergency. Being at a hospital in these times can expose you to COVID-19 and other dangers just by being there. This tool can be used to get a glimpse of information about surrounding hospitals to help make a decision on which one to go if neccesary depending on how busy the hospital beds are.

What it does

This web app lets you search in a US zip code, city or state and renders a map of hospitals in the selected area.

Hospital icons are displayed in the following way:

  • Green: 2/3 or more beds available
  • Orange: Above 1/3 beds available
  • Red: Below 1/3 beds available
  • Gray: If no bed data is available in the dataset

LA Demo Zoomed out

On mouseover, the app displays the percentage of available beds on mouseover, the total number of staffed beds and pediatric and adult ICU beds. On click, the app displays the hospital name and its address.

LA Demo Zoomed in

How I built it

Data description

The dataset is from the AWS Data Exchange and it's called USA Hospital Beds | Definitive Healthcare. Up next is the description given by the data developer:

  • Definitive Healthcare provides intelligence on the numbers of licensed beds, staffed beds, ICU beds, and the bed utilization rate for the hospitals in the United States.

Web App

Using Flask for Python I wrote a web app form that would take in a search inquiry from the user and render another page from there. Then, I used folium to generate maps using the search inquiry. The search inquiry is processed with Pandas since the app looks for the inquiry within a Comma Separated Values (*.csv) dataset.

AWS Resources used

Challenges I ran into

The folium documentation is challenging since it's hard to search in and there are not many examples of fully-fleshed out applications using the framework. Another challenge was using Flask since I had only used it for a brief project before. Most people face the challenge of hosting with AWS, which can be very complicated for someone who has never used AWS before. I have completed some AWS training but never hosted a web app before so I went through the documentation, looked for examples and even contacted AWS help because of a feature I am waiting to hear back so I can implement.

Accomplishments that I'm proud of

Learning how to use two frameworks (Flask and Folium) in a few weeks and being able to integrate them using Object Oriented Programming (OOP) was challenging and rewarding. Being able to deploy to AWS with a custom domain name was very rewarding as well.

What I learned

Web development, AWS Data Exchange, AWS Cloud Services, Flask, OOP, folium, geospatial analysis, Full Stack development, bootstrap, javascript, app forms

What's next for Hospital beds finder

  • Implement the use of geopandas to be able to exploit the geojson dataset version of the data
  • Dailyt bed capacity predictions using machine learning
  • Integrate multiple data sources to make the app not limited to the US
  • Implement "find near me" functionality
  • Add AWS Cloudfront event that triggers AWS Lambda to refresh the dataset S3 with every new revisions by the provider (waiting to hear back from AWS help)

Installation

using a conda virtual environment (optional but recommended)

Naming the environment "geojson"

conda create -n flask python==3.7.6

conda activate flask

conda install jupyter notebook==5.7.8 tornado==4.5.3 (Optional)

Installing dependencies

pip install -r requirements.txt

AWS EB requirements

The following need to be in your requirements file in order for AWS Elastic Beanstalk to serve your app:

  • click==6.7
  • Flask==1.0.2
  • itsdangerous==0.24
  • Jinja2==2.10
  • MarkupSafe==1.1.1 (1.0 can cause errors)
  • Werkzeug==0.14.1

Built With

Share this project:

Updates