Home General How does mapping with demographic shading work?

How does mapping with demographic shading work?

About the Demographic Data

The data comes directly from the U.S. Census Bureau. 

Selections include total population, age group, sex, ethnicity, housing value, household size, number of family/non-family members in household, type of household, household income, education, occupation, etc.

Basic Census Definitions

Source: U.S. Census Bureau and its' affiliate websites

Household

A household includes all the people who occupy a housing unit.  (People not living in households are classified as living in group quarters.) A housing unit is a house, an apartment, a mobile home, a group of rooms, or a single room occupied (or if vacant, intended for occupancy). Separate living quarters are those in which the occupants live separately from any other people in the building, and also, living quarters that have direct access from the outside of the building or through a common hall. The occupants may be a single family, one person living alone, two or more families living together, or any other group of related or unrelated people who share living quarters.

What is the difference between households and families?

A family consists of two or more people (one of whom is the householder) related by birth, marriage, or adoption residing in the same housing unit. A household consists of all people who occupy a housing unit regardless of relationship. A household may consist of a person living alone or multiple unrelated individuals or families living together. 

What is the difference between a median and a mean?

Median income is the amount which divides the income distribution into two equal groups, half having income above that amount, and half having income below that amount. Mean income (average) is the amount obtained by dividing the total aggregate income of a group by the number of units in that group. The means and medians for households and families are based on all households and families. Means and medians for people are based on people 15 years old and over with income.   

Quintiles

  • Think of quintiles like grading on a curve, but instead of a typical bell curve (a few A's, more B's, lots of Cs, some D's and few F's) we are going to grade on a completely FLAT curve.
  • In a class of 100 students taking a test, we are going to FORCE 20 students to get an A, 20 to get a B, 20 get a C, 20 get a D, and 20 students to get an F. . .
  • Even if on a particular test the scores were all grouped together in the high 90%'s or grouped together in the 20%'s or spread out evenly. . . the exact same number of students is going to get each of the 5 grades on that particular test.
  • "Quintile" means "five equal proportions," so the goal of quintile shading is to break a group of data (in this case regions of the earth like DMAs or zip codes) into 5 groups, each group having the same number or count of items in it; more specifically the same count of DMAs or the same count of zip codes. 

 Quintile Shading

  • After we have 5 different groups, we shade the 5 different groups of regions 5 different colors so it is immediately visually obvious which region belongs in which group. 
  • When we grade 100 students, there are 20 students with an A. etc.
  • If we are shading 100 zip codes, there will be 20 zip codes shaded the darkest color (and 20 shaded the lightest color and 20 each of the 3 shades in between).

 Demographic Quintile Shading

  • When we add "Demographic" to make it "Demographic Quintile Shading" we are saying that we are going to grade based on Demographic match to a given target audience. 
  • Our raw scores (the numbers from 0 to 100 for student test scores) are going to be demographics. Specifically, we are going to take the target population of the region (the zip code or DMA) and divide that by the total population. This gives us a percentage number from 0-100 for each region. That is the raw score we are going to force into 5 equal groups and shade. 
  • It could be that every zip code has  less than 5% of their population matching the target audience. It could be that every zip code is going to be over 50% made up of the target audience. In both cases, we are going to make even a subtle difference matter, by forcing the top 1/5 of those zip codes to be shaded the darkest color and bottom 1/5 of those zip codes to be shaded the lightest color. 

 [To strain the grading analogy: Because the regions have different total populations, this is like giving students tests with different numbers of questions. Someone who gets 80% (80) of the questions correct with 100 questions on the test is considered to get the same score as someone that go 80% (800) of the questions correct on a test with 1000 questions] 

DOmedia Demographic Quintile Shading

DOmedia does 2 kinds of Demographic Quintile Shadings.

  1. The first is forced ranking DMAs within the country. Since there are 210 DMAs, for any given target audience there will be 42 DMAs with each of the 5 different shading colors. (210 / 5 = 42) 
  2. The second is forced ranking zipcodes within each DMA. Because different DMAs have different numbers of zipcodes in them, the number of zip codes in each DMA will be different, but in each case, there will be an equal numbers of zip codes in each DMA with each of the 5 shading colors. 

Because we are considering different regions of the earth, the population needed to get the various shading differs depending on which larger region you are looking at. To help you understand this difference while you are looking at a map, DOmedia reports what ratio of target audience to total audience it took to get each shading level; what percentage of right answers on the test it took to get an A, B, C, D and F. . .  Based on the distribution of test scores or based on the target population numbers, they are going to be different and may be widely varying numbers. Because the populations are different, we display the %'s needed to get each shading. This is exactly like a professor posting the results of a test with "it took a 78% right to get an A on this test." We are saying "It took 5% of the population matching the target audience to get the darkest shading, for this target audience, for this region" 

Detailed Caveats and Special Cases

Some details I left out in the description that are technically true, but would have complicated the explanation unnecessarily:

  • When the number of regions doesn't evenly divide by 5, we make the darkest region have the extra highest scoring regions (So if we have 28 zip codes in a DMA, we'd have 5 zipcodes the lightest shading, 5 the darker ,5, 5 and then 8 shaded the darkest color to add up to 28. With 29 zip codes we'd group 5,5,5,5,9. With 30 zip codes we'd group 6,6,6,6,6).
  • A score of 0 always results in the lightest shading, even if that means the lightest shading "bucket" has more than 1/5 of the total number of zip codes in it.
  • For shadings that ARE the total population (total population, total households, etc) because there is nothing to divide by, we use the population as the score and mark the legend with the raw numbers instead of the percentages. 
  • Because a DOmedia map can be arbitrarily zoomed in or out, we always shade the zip codes within the DMA they are in, even if we are displaying multiple DMAs at the same time. This has the advantage of the shading being consistent independent of zoom level, but has the disadvantage of the view not technically, strictly, quintile shaded. Please let DOmedia know if you would like to discuss this nuance in detail. 
  • The actual definition of "Quintile" is "A Quantile for the special case of five equal proportions" and the definition of a "Quantile" is "one of the class of values of a variate that divided the total frequency of a sample or population into a given number of equal proportions" (aren't you glad I left that out?)