1. Acquiring Population Data
The first step in any RSTbx workflow is a boundary file and population data.
Luckly, the Census Data Retriever can do both of those things for us!
Requirements
Boundary file
The RSTbx supports any geographic unit for which you can provide a boundary file for. Typically, people utilize Census boundary files for counties or tracts, which can be downloaded in TIGER or Cartographic forms from the Census website.
There are three requirements of these boundary files:
- They should specify polygon/multipolygon boundaries.
- There must be at least one column which must uniquely identify each region within the boundary file. These unique identifiers must be the same as those within the population table and the event table.
- There should be no geographic regions within the boundary file which are not adjancent to at least one other area.
Using Census TIGER boundaries rather than cartographic boundaries will mitigate the concern of geographic units without any adjacent areas, because TIGER boundaries represent political boundaries which span over water.
Population Tables
Population tables can come in two formats depending on your eventual goals. If you would like to produce crude rates, then the table should have one record (row) for each region. It would look something like this:
| GEOID | PopulationCount |
|---|---|
| 01001 | 59285 |
| 01003 | 239945 |
| 01005 | 24757 |
| … | … |
| 72149 | 21778 |
| 72151 | 29868 |
| 72153 | 33509 |
(Also see MI_pop within the data.gdb)
If you would like to produce age-adjusted rates, then the table should have one record for each region age-group combination. It would like something like this:
| GEOID | AgeGroup | PopulationCount |
|---|---|---|
| 01001 | 0-4 | 3430 |
| 01001 | 5-14 | 7749 |
| 01001 | 15-24 | 7339 |
| … | … | … |
| 72153 | 65-74 | 4527 |
| 72153 | 75-84 | 3223 |
| 72153 | 85up | 571 |
(Also see MI_pop_grouped within the data.gdb)
The following are valid age groups: “0-4”, “5-14”, “15-24”, “25-34”, “35-44”, “45-54”, “55-64”, “65-74”, “75-84”, “85up”. Within the population table they must be written exactly how they are listed here (i.e., “85 and up” would not be a valid age group).
Using custom data
Being able to use custom data within the RSTbx is a great asset.
One potential use of custom data within the RSTbx is to calculate rates at custom geographic regions. For example, if I wanted to calculate the heart disease death rate at Texas Public Health Regions, I could use a custom boundary file and population table to do so!
Another potential use is to calculate group-specific rates. For example, if I desired to calculate the heart disease death rate for females, I could filter down my event data to just female deaths and use a custom population table with total female population.
The RSTr package has additional models which are more well suited to calculating group-specific rates. We encourage you to check the RSTr out if you are looking to calculating group-specific rates.
All the requirements above for your boundary file and population table will still hold even if you choose to use custom boundaries or population tables.
Usage
All succeeding tutorials will rely on custom data provided within the data.gdb, but this tutorial will help you become familiar with the CDR. You will not use this data in any of the other tutorials.
If you are looking to generate rates at the county (or county-equivalent) or tract level, the Census Data Retriever can prepare your data for you.
As discussed above, two pieces of data must be acquired as part of the first step of an RSTbx workflow: a boundary file and a population file.
Our eventual goal is to produce both crude and age-adjusted rates, so we will require a boundary file for both, a total population table, and an age-stratified table.
Producing crude rates
In order to create our crude rates, we will need a total population table. Let’s create one for Michigan counties using the 2015-2019 5-year ACS.
If you haven’t already, set up the Rate Stabilizing Toolbox.
Open the Census Data Retriever
Set the following parameters and Run:
Age-Stratified: Unchecked (We only need total population numbers)
Request Parameters:- Survey: 5-year ACS
- Year: 2015-2019
- Geography: County
- State: Michigan
Output Table: mi_county_acs1519_pop
- Survey: 5-year ACS
Producing age-adjusted rates
Now that we have downloaded our total population table. We still require a age-stratified population table to calculate age-adjusted rates. Additionally, we still require boundaries. Fortunately, we can take care of both of these in one step.
Returning back to the Census Data Retriever, set the following parameters:
Age-Stratified: Checked (We want an age-stratified table)
Keep all the Request Parameters the same:- Survey: 5-year ACS
- Year: 2015-2019
- Geography: County
- State: Michigan
Output Table: mi_county_acs1519_pop_grouped
We also want to download our boundaries, so open up the Geography dropdown.
Set the Geometry Type to Cartographic and the Output Feature to mi_county_carto. Run.
- Survey: 5-year ACS
We have seen how the Census Data Retriever can be used to prepare population data, but we still need to prepare our event data. If you are interested in using individual, record level event data, consider moving on to 2a. Preparing Individual-Level Event Data. If you are interested in using aggregate event data, consider moving on to 2b. Preparing Aggregate Event Data.