---
tags: Ongoing Experiments
---

# Experiment 6: Aligning Sim to Real QC Data

### Description
Here we have [real-world QC data](https://covid19tracker.ca/) and simulated data (in this case using a population of 1,000 people with initial percentage sick of 0.004%) and GLOBAL_MOBILITY_SCALING_FACTOR of 0.1

### Quebec Data

### Simulated Data (10k people, 50 people initially sick, mobility = 0.02)

### Simulated Data (10k people, 50 people initially sick, mobility = 0.05)

### Simulated Data (10k people, 50 people initially sick, mobility = 0.1)

### Simulated Data (10k people, 50 initially sick, mobility = 0.15)

### Simulated Data (10k people, 50 initially sick, mobility = 0.2)

### Tests
Is quebec testing at capacity right now? How should we align our simulated testing with this given that many times more people than can be tested want a test in our sim.

### Analysis
- higher percent initially sick results in faster growth in cases, mortalities, and hospitalizations
- higher mobility also results in faster growth in cases/mortalities/hospitalizations
- with all variables remaining consistent, and only the population size changing, the growth rates and trends of cases/mortalities/hospitalizations are similar if not the same
- need to compare first 30 days of Quebec data to simulated 30 days for more insights

### Outcomes / Next Experiments
- plot everything cumulatively (or not cumulatively)
- compare first 30 days of quebec data against sim data
- determine what the largest simulation we can run is (population size)
- add caption percent population => to number of initial people sick
- goodness of fit test
- tune parameters
- reduce init percent infected, reduce GLOBAL_MOBILITY_SCALING_FACTOR

### backlog:
- add dynamic restrictions to align with qc data ()
- Look more closely at [JPC's spreadsheet](https://docs.google.com/spreadsheets/d/16uOJXX4VVHBICqk1M4LeTYI9Q0oT6rJsvrO2kZQLFsE/edit#gid=0)
- check if hospitalization data is reliable