Exploring Summary Statistics Using Google Sheets & Data Wrapper

Data Wrapper – Example #1:

Figure 1 below illustrates toxic chemical releases by state, reported to the EPA Toxics Release Inventory in 2009. The data were distributed in the US Census Bureau in the US Statistical Abstracts for 2012.

Figure 1. US EPA Toxics Release Inventory total by US State, 2009, reported in millions of pounds

As shown in Table 1, I found that the average across U.S. states and DC was 66 million lbs. The highest value recorded was for Alaska, which released nearly 700 million lbs. of toxic chemicals in 2009. Total releases for the United States were 3.37 billion lbs.

Data Wrapper – Example #2:

Figure 1 below illustrates marijuana use by state in 2007-2008, with data reported in the US Statistical Abstracts and collected by CDC surveys.

Examples with Google Sheets:

Google Sheets, which is the online spreadsheet application provided by Google, enables users to create maps as charts within spreadsheets.  This is a very useful way to explore summary statistics.  Four examples are shown below.

Example One: Distribution of Working Age Veterans in the United States

Figure 1.  Distribution of US Veterans, 18-64 years old, as a Percent of Total Population: American Community Survey, 2013-2017, downloaded as tables from Social Explorer.

Table 1, below, summarizes the statistical findings from the American Community Survey (2013-2017).  The average percentage of working age veterans is around four percent.

Last, figure 2 illustrates the distribution of working age veterans in a histogram.

Figure 2. Histogram of Veteran Populations by State and Territory, American Community Survey (2013-2017)

The data for my example can be found here.

Example Two: Unemployment in the U.S.

Figure 1. Unemployment Rate by U.S. State, Reported in the 2012-2016 Cumulative American Community Survey (ACS), Distributed by Social Explorer.  Interactive map available online.

Table 1 describes the summary statistics for employment.  The rates of unemployment ranged from 2.77 percent in North Dakota to a high of 17.7 percent in Puerto Rico.  The median value, 7.16 percent, was found in between Delaware (7.13 percent) and Pennsylvania (7.19 percent).

The histogram shown in Figure 2 highlights the extent to which Puerto Rico is an outlier against all other measured unemployment rates for U.S. states.

Figure 2.  Histogram of Unemployment across U.S. States

The original data and charts can also be found on Google Sheets, which is shared here.

Example Three: Per Capita Income in the U.S.

Figure 1 illustrates the distribution of per capita income across U.S. states.  The highest incomes were in the Northeast and the lowest in the Southeast.

Figure 1.  Per Capita Income across U.S. States, ACS (2012-2016).  Interactive map available online.

Table 1 describes the summary statistics for Per Capita Income.  The average income for the states was $29,332, with the highest found in Connecticut ($39,906) and the lowest in Mississippi ($21,651).  The median value lies between Kansas and Maine, at $28,476.

Figure 2 below shows the histogram, or distribution of per capita income, across U.S. states.  The distribution is somewhat normal across states.

Figure 2.  Histogram of Per Capita Income across U.S. States

The data and charts presented here can also be found online.

A visualization of the ACS data is also available through Social Explorer, as seen below.

Maps in Google Sheets may also be embedded directly onto your web page, which creates an interactive visualization.

Example Four: Income Inequality as Shown by the Gini Coefficient

Figure 1 below shows income inequality by state.  The highest inequality is found in New York and the lowest in Alaska.  The Southeastern states show higher inequality than the Great Plains and Midwestern states.

Figure 1. Gini Coefficient across States, American Community Survey (2013-2017)

Table 1 shows the summary statistics for the Gini Coefficient from the American Community Survey, 2013-2017.

Figure 2. Histogram of the Gini Coefficient, American Community Survey (2013-2017)

Figure 2 above shows the frequency distribution of the Gini Coefficient by state. It appears to be a normal distribution.

The data tables for the charts above can be found here.

Example 5: Comparing Black and Mainline Protestant Populations in the US

Figure 1. Mainline Protestant Populations by State (2010)

Figure 2. Black Protestant Population by State (2010)

Figure 3. Scatterplot Illustrating the Correlation between Black Protestant Adherence Rates and Mainline Protestant Adherence Rates