Analyzing Monthly Household Survey with Shiny

  1. Employment Situation Summary
  2. Reproducing Numbers from the Employment Situation Summary in Shiny
  3. Graphing the Unemployment Rate over Time
  4. Graphing the Change in the Counts over Time, by Employment Status

Employment Situation Summary

The Employment Situation Summary is described at this link as follows:

The U.S. Bureau of Labor Statistics (BLS) releases the Employment Situation Summary, better known as the employment, or jobs report, at 8:30 a.m. ET on the first Friday of every month. The report is based on surveys of households and employers. It estimates the number of people on payrolls in the U.S. economy, the average number of hours they worked weekly, and their average hourly earnings, along with several versions of the unemployment rate.

The most recent Employment Situation Summary can be found at https://www.bls.gov/news.release/empsit.nr0.htm. Following the summary are links to Employment Situation Summary Table A. Household data, seasonally adjusted and Employment Situation Summary Table B. Establishment data, seasonally adjusted. The meaning of "seasonally adjusted" is explained at this link. Following the links to these two summary tables are links to detailed tables A-1 through A-16 for the Household Survey data and tables B-1 through B-9 for the Establishment Survey data. Following is the beginning of Table A-1:

Note that columns 2 to 4 contain numbers that are "Not seasonally adjusted" and columns 5 to 10 contain numbers that are "Seasonally adjusted".

Reproducing Numbers from the Employment Situation Summary in Shiny

Many of the "not seasonly adjusted" numbers in Table A-1 can be reproduced via the publicly available Shiny app at https://econdata.shinyapps.io/cps_monthly1/. Following is the initial page that comes up when that app is accessed:

Following is the table in the above page with columns added for corresponding values in Table A-1 above:

All Persons: 2023-2024, grouped by EMPSTAT, AGE 16-99 (percent)

                          Table A-1
                           Civilian                        Table A-1
                        Labor Force                     Unemployment
   Year_Mo       Count  (thousands) Employed Unemployed         Rate
1  2023-01 165,069,905                  96.1        3.9
2  2023-02 166,177,943                  96.1        3.9
3  2023-03 166,783,207      166,783     96.4        3.6          3.6
4  2023-04 166,220,828                  96.9        3.1
5  2023-05 166,702,251                  96.6        3.4
6  2023-06 167,910,035                  96.2        3.8
7  2023-07 168,353,639                  96.2        3.8
8  2023-08 168,049,264                  96.1        3.9
9  2023-09 167,718,080                  96.4        3.6
10 2023-10 167,774,401                  96.4        3.6
11 2023-11 167,976,734                  96.5        3.5
12 2023-12 166,660,598                  96.5        3.5
13 2024-01 166,427,532                  95.9        4.1
14 2024-02 167,285,357      167,285     95.8        4.2          4.2
15 2024-03 167,960,219      167,960     96.1        3.9          3.9
As can be seen, the 3 Civilian labor force numbers in the not seasonally adjusted columns of Table A-1 match the generated Count numbers, rounded to the nearest thousands. Also, the 3 Unemployment rate numbers from Table A-1 match those in the generated Unemployed column.

The Shiny app generates these numbers from the monthly data files at this link using the Data Dictionary linked to at the top of that page. The data for each month is usually available within a week of the release of that month's job report.

Changing the Units select list to Count, the Employment status select list to All, and clicking the Submit button will result in the following page:

All Persons: 2023-2024, grouped by EMPSTAT, AGE 16-99 (count)

                          Table A-1                Table A-1               Table A-1                       Table A-1
                           Civilian
			Noninstitut.                                                                          Not in
                         Population                 Employed              Unemployed                     Labor force
   Year_Mo       Count  (thousands)    Employed  (thousands) Unemployed  (thousands) Not.in.labor.force  (thousands)
1  2023-01 265,961,720              158,691,547               6,378,358                     100,891,815             
2  2023-02 266,111,907              159,712,565               6,465,378                      99,933,964             
3  2023-03 266,271,876      266,272 160,740,630      160,741  6,042,577      6,043           99,488,669       99,489
4  2023-04 266,443,193              161,074,745               5,146,084                     100,222,365             
5  2023-05 266,618,231              161,002,386               5,699,864                      99,915,980             
6  2023-06 266,800,905              161,559,139               6,350,896                      98,890,870             
7  2023-07 267,002,315              161,981,640               6,372,000                      98,648,676             
8  2023-08 267,213,250              161,426,705               6,622,559                      99,163,986             
9  2023-09 267,427,697              161,669,304               6,048,777                      99,709,617             
10 2023-10 267,642,080              161,676,321               6,098,081                      99,867,679             
11 2023-11 267,822,020              162,149,298               5,827,436                      99,845,286             
12 2023-12 267,991,084              160,753,981               5,906,617                     101,330,486             
13 2024-01 267,540,215              159,649,766               6,777,766                     101,112,683             
14 2024-02 267,711,141      267,711 160,315,160      160,315  6,970,197      6,970          100,425,784      100,426
15 2024-03 267,884,190      267,884 161,356,417      161,356  6,603,802      6,604           99,923,971       99,924
As can be seen, the 3 Civilian noninstitutional population numbers in the not seasonally adjusted columns of Table A-1 match the generated Count numbers, rounded to the nearest thousands. This also holds true for the 3 Employed, Unemployed, and Not in labor force numbers.

Graphing the Unemployment Rate over Time

The unemployment rate over time can be graphed via the following steps:

  1. Go to https://econdata.shinyapps.io/cps_monthly1/.
  2. Set First Year to 2015.
  3. Set Min data to 3. This will cause the table to display only the third numeric column, Unemployed.
  4. Click the Submit button. This will cause the data from 2015 to 2022 to be loaded and displayed.
  5. Click the Plotly tab.
  6. Hover the cursor over the rightmost point in the graph. That will display the following page:

Note that the unemployment rate generally trended down from 2015 to February of 2020 when it then spiked up to a maximum in April of 2020 due to COVID. It then declined by 2023 to its pre-COVID low. Also, note that the unemployment rate tends to go up around January and June of each year. This appears to be the major seasonal pattern and is likely not visible in the seasonally adjusted data.

Regarding seasonal adjustments, the following message is posted by a former IPUMS staff member on an IPUMS forum:

This article goes into great detail on the seasonal adjustment procedures used for labor force statistics by the BLS. Starting with IPUMS CPS microdata, you'll first need to calculate the aggregate statistics you are interested in by month, and then run the whole time series through a seasonal adjustment program such as X-13 ARIMA SEATS.

Hence, seasonally adjusting the data is not a trivial operation and this Shiny app currently does not perform that adjustment. On the other hand, the Employment Situation Summary does usually refer to the seasonally adjusted data. For example, the first line in the March 2024 report states that "the unemployment rate changed little at 3.8 percent". The 3.8 percent matches the seasonally adjusted unemployment rate in Table A-1 whereas the 3.9 percent for the rightmost point in the graph above matches the "not seasonally adjusted" employment rate in that table.

Graphing the Change in the Counts over Time, by Employment Status

  1. Go to https://econdata.shinyapps.io/cps_monthly1/.
  2. Set First Year to 2015.
  3. Set Units to Count.
  4. Set Employment status to All.
  5. Set Min data to 1. This will cause the table to display all numeric columns.
  6. Check the View Change checkbox next to the Submit button.
  7. Click the Submit button. This will cause the data from 2015 to 2022 to be loaded and displayed.
  8. Click the Plotly tab.
  9. Set Y Units to millions on the second leftmost side panel. That will display the following page:


Analyzing Employment Data from the Establishment and Household Surveys
Go to Budget Home Page

free
web stats