Most visualizations contain at least one calculated value to help you understand the data you see. The three values that Rhiza for Research calculates are:
- Proportion: shows the distribution of data within a population
- Percentage: compares two populations and shows how they differ; useful for things like market share and trends
- Index: compares two populations and shows how they differ; useful for things like market share and trends
Proportion is the most common calculated value you'll see in a table; it is calculated by dividing the number of respondents (or car registrations, or other data point) for each data group by the overall number of respondents (or car registrations, or data points) in the data series. The individual proportions must always add up to 100%.
If it helps, you can think of proportion as a fraction, with the numerator as the individual number of respondents in your query and the denominator as the entire set of respondents in the data series.
The following table shows two proportions: one for the raw number of respondents and one for the weighted number of individual respondents (this particular dataset models the data to project numbers that represent the given market). In this example, we defined a data series that returned women in the New York DMA who dined out, and grouped the responses by income level. When you divide the respondent count for an income level by the net total, you get the proportion. It essentially says, "Out of all the women diners in New York, 15.3% of them make between $50,000 and $74,999 per year."
Figure 41: An example of a table showing proportion (click to enlarge)
The index is used as a comparison tool. As such, it requires you to create and use a context series so that you are comparing two populations: the target demographic in your data series, and the larger universe in your context series.
It is calculated by dividing the proportion by the value defined by that context series. A value of 100 typically means that your target demographic or product is doing well compared to some larger entity; values higher than 100 mean that it's doing very well comparatively.
Let's go back to the example we used for proportion. We had a data series that examined women diners in the New York DMA, and then grouped the results by income level. To calculate the index, we need to add a context series that encompasses some larger set of respondents. In this case, we'll add a context series that includes all diners nationwide. Our resulting table now includes an index value.
Remember that one of the largest proportions of diners made between $50,000 and $74,999 per year. However, when we look at the index, we see that, when compared to all diners nationally, they under-index. That is, that demographic in New York eats out less frequently when compared to the rest of the dining-out nation. However, women in New York who make between $150,000 and $249,000 per year are almost 1.4 times more likely to eat out than the average national diner in that income bracket.
Figure 42: A table showing the calculated proportion and index values (click to enlarge)
You can also show the index as a line on a bar or column chart to help you visualize it.
Figure 43: A column chart showing the calculated index value
Percentage is another comparison tool. As such, it also requires you to create and use a context series so that you are comparing two populations: the target demographic in your data series, and the larger universe in your context series.
It is calculated by dividing the value for the data in each grouping by the context series value.
Because this calculation is so useful for things like market share, let's use a slightly different example to illustrate it. We've defined a data series to examine soda drinkers in the New York DMA, and grouped them by the brand of soda they drink most often -- we want to know what brands have the most market share. We've also defined a context series that includes all soda drinkers nationwide. Our resulting table looks like the following.
Figure 44: A table that shows the calculated percentage when used with a context series (click to enlarge)
When we divide the number of responses for Coca Cola by the number given in the context series, we see that 5.5% of the weighted individual respondents in New York prefer Coca Cola; that's its market share, essentially. Note that the percentage in the Net Total row does not equal 100. This calculation is showing the market share for the defined demographic -- in this case, soda drinkers in the New York DMA. We should expect this percentage to be relatively small when compared with the context series (in this case, the nation) -- it makes sense that New York soda drinkers account for only 6.8% of all soda drinkers in the United States.
© 2016, Rhiza, Inc. All rights reserved. Last updated June 06, 2016 03:10:52 PM.