1 of 8

Ontario Swift ER

Abdul Rahman Mousa & Nhat Truong Luu

2 of 8

Theme

3 of 8

Problem

  • Hospital wait times in Canada
    • Wait time to be assessed by a doctor in the ER
  • Last year ~500,000 patients left the ER without seeing a doctor - CBC
  • Closest option is not always the best option

​

4 of 8

Solution/Impact

  1. Show users where to get the quickest initial assessment
    • Travel time + Wait time taken into account
  2. Impact:
    • Reduce wait time for patient
    • Ease stress of patient
    • Even out the workload on hospitals

​

5 of 8

Execution

​

Tech Stack:

  • Streamlit
  • Selenium
  • Jupyter notebooks
  • Google Cloud APIs: Geocoding API + Distance Matrix API
  • Please refer to requirements.txt to see all dependencies used and their versions

6 of 8

Execution

Data Sources:

  1. Web scraped data set Health Quality Ontario
    • For hospital wait time + length of stay stats across ontario
    • Data Dictionary
  2. Ontario Geo Hub
    • Geographic data (long,lat,address) on hospitals
  3. Merged datasets by hospital name

7 of 8

Execution cont.

​

4) Data Science web app

  • User inputs:
    • Address
    • Max time willing to wait
  • ML Model ranks 5 hospitals with shortest waiting time
  • Displays
    • Map visualization of Hospitals
    • Details to user

8 of 8

Machine Learning

KNN MODEL

Distance

Total Time

(Travel + Waiting)

KNN