Equity Metric

Equity Metric#

Description

The algorithm computes equity metrics to characterize the inequity present in the infrastructure service provision between two groups of concerns’ (e.g., low income vs. non-low income, minority vs non-minority, etc.). The metrics are built upon Theil’s T, which is a common metric to compute the inequality present in the dispersion of a scarce resource (e.g., income). The metrics could also be implemented to assess inequity in different scenarios with other scarce resources of concern.

To compute the metrics, a scarce resource should be formulated and calculated for an infrastructure application. The scarce resource for infrastructure can be taken as any user defined relevant values, such as a resilience score (i.e., probability of service provision) or recovery time. We’ve included a companion utility class where we define and prepare recovery time as a scarce resource. Other scarce resources can and could be explored if provided by the user.

The equity metric allows for an equity assessment of the current infrastructure service provision. It also enables the assessment of equity gains for a given retrofit plan and can be integrated into an overall decision-making process.. The output metrics tell the following information 1) Theil’s T - overall amount of inequality in scarce resource’s dispersion across a community (distributional inequity) 2) Between Zone Inequality (BZI) - amount of inequality attributed to scarce resource differences between groups (restorative inequity), and 3) Within Zone Inequality (WZI) - amount of inequality due to resource differences among singular groups.

Contributors

Science: Abigail L. Beck, Ph.D, Eun Jeong Cha, Ph.D, Walter Peacock, Ph.D
Implementation: NCSA IN-CORE Dev Team

Related publications

Beck, A.L., Cha, E.J. & Peacock, W.G. “Incorporation of Equity into Infrastructure Decision-Making: Development of an Equity Metric for Infrastructure Retrofitting,” The 14th International Conference on Applications of Statistics and Probability in Civil Engineering (ICASP14), Dublin, Ireland, July, 2023. http://hdl.handle.net/2262/103309

Input parameters

key name	type	name	description
`result_name` ^*	`str`	Result name	Name of the result dataset.
`division_decision_column` ^*	`str`	Division decision column name	Column name of a binary variable associated with each household used to group it into two groups (e.g. low income vs non low income, minority vs non-minority, social vulnerability)

Input datasets

key name	type	name	description
`housing_unit_allocation` ^*	`incore:housingUnitAllocation`	Housing allocation	A housing unit allocation dataset.
`scarce_resource` ^*	`incore:scarceResource`	Scarce resource	Scarce resource dataset e.g. probability of service, return time, etc.

Output datasets

key name	type	parent key	name	description
`equity_metric` ^*	`incore:equityMetric`		Equity Metric	CSV file of equity metric, including Theil’s T Value, Between Zone Inequality, Within Zone Inequality

(* required)

Execution

code snippet:

    client = IncoreClient()
    datasvc = DataService(client)

    # Example of preparing scarce resource
    repair_time_df = Dataset.from_data_service(housing_recovery_id, datasvc).get_dataframe_from_csv()
    scarce_resource_df = EquityMetricUtil.prepare_return_time_as_scarce_resource(
        repair_time_df
    )
    scarce_resource = Dataset.from_dataframe(
        scarce_resource_df, "scarce_resource", data_type="incore:scarceResource"
    )
    
    # Example of running equity metric analysis
    equity_metric = EquityMetric(client)
    equity_metric.set_parameter("result_name", "Galveston_recovery_time")
    equity_metric.set_parameter("division_decision_column", "SVI")
    equity_metric.load_remote_input_dataset(
        "housing_unit_allocation", housing_unit_allocation_id
    )
    equity_metric.set_input_dataset("scarce_resource", scarce_resource)
    equity_metric.run_analysis()

full analysis: equity_metric.ipynb