Adding age-demographics to pipeline

[rathp]

11

Adding Counts in Age Group for Every Stage and Time-step

Summary

• Updated param_simple_example.json, PatientTrajectory.py and run_forecast to now simulate counts for 4 age groups : 0-19", "20-44", "45-64", "65+.

• These specific age groups were set as per literature survey done here : [https://docs.google.com/document/d/1a1lKY4ElVh7bh9RpD7TL3kzhfZGn13ELtsZRlUYCNsM/edit?usp=sharing]()

Implementation

• In each forward run of the forecast, each person get's assigned an age_group attribute depending on the probabilities specified in proba_Presenting_given_* defined in the parameter json file.

• Along with the counts in each state, the counts in each group and each state are also computed as update_*_matrix_by_age() methods in the PatientTrajectory class.

Example simulation

• python run_forecast.py --config_file "params_simple_example.json" --random_seed 8675309 --output_file "results.csv"
• Output :

  ----------------------------------------
  Simulating for 120 timesteps with seed 8675309
  ----------------------------------------
  100%|████████████████████████████████████████████████| 120/120 [00:00<00:00, 301.19it/s]
  ----------------------------------------------
  Printing Age Group Counts for 5 time steps
  ----------------------------------------------
  ----------------------------------------------
  Time Step : 0
  ----------------------------------------------
  {'occupancy_0-19': array([0., 0., 1., 0.]), 'occupancy_20-44': array([0., 3., 1., 1.]), 'occupancy_45-64': array([0., 3., 5., 4.]), 'occupancy_65+': array([0., 4., 3., 5.]), 'occupancy_total': array([ 0., 10., 10., 10.])}
  ----------------------------------------------
  Time Step : 1
  ----------------------------------------------
  {'occupancy_0-19': array([0., 0., 1., 0.]), 'occupancy_20-44': array([1., 2., 1., 1.]), 'occupancy_45-64': array([2., 3., 4., 3.]), 'occupancy_65+': array([6., 2., 4., 4.]), 'occupancy_total': array([ 9.,  7., 10.,  8.])}
  ----------------------------------------------
  Time Step : 2
  ----------------------------------------------
  {'occupancy_0-19': array([0., 0., 1., 0.]), 'occupancy_20-44': array([0., 3., 0., 1.]), 'occupancy_45-64': array([4., 1., 4., 4.]), 'occupancy_65+': array([5., 4., 5., 3.]), 'occupancy_total': array([ 9.,  8., 10.,  8.])}
  ----------------------------------------------
  Time Step : 3
  ----------------------------------------------
  {'occupancy_0-19': array([0., 0., 0., 1.]), 'occupancy_20-44': array([2., 0., 3., 1.]), 'occupancy_45-64': array([5., 5., 2., 5.]), 'occupancy_65+': array([6., 7., 4., 5.]), 'occupancy_total': array([13., 12.,  9., 12.])}
  ----------------------------------------------
  Time Step : 4
  ----------------------------------------------
  {'occupancy_0-19': array([0., 0., 0., 1.]), 'occupancy_20-44': array([0., 2., 2., 2.]), 'occupancy_45-64': array([6., 7., 4., 4.]), 'occupancy_65+': array([ 4., 10.,  3.,  4.]), 'occupancy_total': array([10., 19.,  9., 11.])}
  ----------------------------------------

• For each time-step in the output above, the occupancy_total always adds up the sum of occupancy_* across age groups.