Household-level housing sequential recovery

Household-level housing sequential recovery#

Description

This analysis computes the series of household recovery states given a population dislocation dataset, a transition probability matrix (TPM) and an initial state vector.

The computation operates by segregating household units into five zones as a way of assigning social vulnerability. Using this vulnerability in conjunction with the TPM and the initial state vector, a Markov chain computation simulates most probable states to generate a stage history of housing recovery changes for each household.

The output of the computation is the history of housing recovery changes for each household unit in CSV format.

Contributors

Science: Elaina Sutley, Sara Hamideh
Implementation: Nathanael Rosenheim, Santiago Núñez-Corrales, and NCSA IN-CORE Dev Team

Related publications

Sutley, E.J. and Hamideh, S., 2020. Postdisaster housing stages: a Markov chain approach to model sequences and duration based on social vulnerability. Risk Analysis, 40(12), pp.2675-2695.

Input parameters

key name	type	name	description
`result_name` ^*	`str`	Result name	Name of the result dataset.
`seed` ^*	`int`	Seed	Initial value to seed the random number generator to ensure replication of the Markov Chain path’ in connection with Population Dislocation.
`t_delta` ^*	`float`	Time step	A size of the analysis time step.
`t_final` ^*	`float`	Time duration	Total duration.

Input datasets

key name	type	name	description
`population_dislocation_block` ^*	`incore:popDislocation`	Population dislocation	Population dislocation results.
`tpm` ^*	`incore:houseRecTransitionProbMatrix`	Probability matrix	A transition probability matrix that specifies the corresponding Markov chain per social vulnerability level.
`initial_stage_probability` ^*	`incore:houseRecInitialStageFactors`	Mass probability	Initial mass probability function for stage 0 of the Markov Chain.

Output datasets

key name	type	parent key	name	description
`result` ^*	`incore:housingRecoveryHistory`	`housing_recovery_block`	Results	A dataset containing results (format: CSV) with housing recovery sequences at the individual household level.

(* required)

Execution

code snippet:

    # Create Housing recovery sequential analysis instance
    hrs = HousingRecoverySequential(client)

    # Load input dataset
    hrs.load_remote_input_dataset("population_dislocation_block", population_dislocation_block)
    hrs.load_remote_input_dataset("tpm", tpm)
    hrs.load_remote_input_dataset("initial_stage_probability", initial_stage_probability)

    # Specify the result name
    result_name = "IN-CORE_housingrecovery"

    # Set analysis parameters
    hrs.set_parameter("result_name", result_name)
    hrs.set_parameter("seed", 1238)
    hrs.set_parameter("t_delta", t_delta)
    hrs.set_parameter("t_final", t_final)

    # Run Housing recovery analysis
    hrs.run_analysis()

full analysis: housing_household_recovery.ipynb