Creating a Survey

Overview

The Survey Tool supports the user in performing a survey of structures in an NSI inventory. There are two primary ways to define a new survey:

Create Survey from CSV
Generate Stratified Survey

Both methods create a new survey that contains the following primary elements:

A survey name and description
A set of members and owners
A survey polygon location
A set of survey elements containing structure identifiers, strata names, and if an element is a control or not.

Creating a New Survey

To create a new survey, select the "Create New Survey +" button in the upper left corner of the dashboard page.

This will change the Active Surveys display to the Create New Survey display

Provide a Survey Name, Due Date, and Description. By default the person creating the new survey will be automatically added as the owner of the survey. A survey must have at least one member who is the owner. The list of users is provided to select additional survey members from. Multiple members can be owners of a survey. A survey owner can edit, delete, and view for a survey they own.

Adding Survey Elements

There are two options for adding survey elements, user uploaded CSV file or "Generate Stratified Survey". Both methods require the user to select the inventory source and to upload a study area polygon as a geojson file with a single polygon feature.

User Uploaded CSV

The "Generate Stratified Survey" checkbox is defaulted to false, which means the default way to add survey elements is by user uploaded list. This method requires the user to upload a csv with the columns fd_id, strata, and control.

For more information on how to create a CSV see the section on "Creating a Sample list"

Generate Stratified Survey

Generating a Stratified survey relies on the user selected inventory of choice, the user provided study area boundary and the type of strata to be applied. The user is given a choice between stratification by single family residential and all other structures and/or FEMA flood zone A and V and all other zones.

A survey must have a user defined confidence, and margin. Each strata also needs a proportion. The margin, confidence and proportion (by strata) are used to define the survey sample size using Cochran's formula to determine sample size:

$\begin{array}{l}\displaystyle n_0 = \frac{Z^2 p(1-p)}{e^2}\end{array}$

The finite population adjustment is also utilized:

$\begin{array}{l}\displaystyle n = \frac{n_0}{1 + \frac{n_0 - 1}{N}}\end{array}$

Where:

Z = The z-score for the desired confidence level (e.g., 1.96 for a 95% confidence level)
p = The estimated proportion of an attribute present in the population (use 0.5 for the most conservative sample size)
e = The acceptable margin of error (e.g., 0.05 for ± 5%)
N = The total population size of the strata
n₀ = The sample size for each strata

The full sample size is the sum of all strata sizes.

Finally a percent control structure is necessary to define a survey, the first 10 structures are all control structures for every generated survey and then all remaining structures have the user defined percentage of structures randomly selected as control structures.

When the user clicks the button generate stratified survey the NSI API is queried for all structures in the user provided study boundary, structures are sorted into their strata and counted, Cochran's formulas are applied and sample sizes are estimated for each strata. Then structures are randomly assigned up to the sample size from each strata and assigned as control structures. A report showing the first few sampled survey elements is provided to the user and then the user can continue to "Create Survey".

Caution

If a user changes settings such as the margin, confidence, or proportion by strata then the blue "Generate Stratified Survey" button must be clicked again to respect the user's changed inputs.