Residential Quota Downselect Sampler

Sometimes it is desirable to run a stock simulation of a subset of what is included in a project. For instance one might want to run the simulation only in one climate region or for certain vintages. However, it can be a considerable effort to create a new project and modify the housing characteristic distributions. The Residential Quota Downselect sampler adds a downselection capability to the Residential Quota Sampler.

Downselecting can be performed in one of two ways: with and without resampling. Downselecting with resampling samples twice, once to determine how much smaller the set of sampled buildings becomes when it is filtered down and again with a larger sample so the final set of sampled buildings is at or near the number specified in n_datapoints.

Downselecting without resampling skips that step. In this case the total sampled buildings returned will be the number left over after sampling the entire stock and then filtering down to the buildings that meet the criteria.

Configuration Example

sampler:
  type: residential_quota_downselect
  args:
    n_datapoints: 350000
    logic:
      - Geometry Building Type RECS|Single-Family Detached
      - Vacancy Status|Occupied
    resample: false

Arguments

  • n_datapoints: The number of datapoints to sample.

  • logic: The criteria to apply to remove buildings from the simulation. For details on how to specify the filters, see Filtering Logic.

  • resample: boolean, whether to run the sampling twice with a goal to have the downselected number of datapoints be equal to n_datapoints.