SIMPLE ARIMA FORECAST OF BLS NONFARM PAYROLLS.
This script (and accompanying data file) produces auto-ARIMA forecasts of BLS NSA payrolls, as well as implied BLS seasonal adjustment factors. Combines into a 12-month-ahead forecast of monthly SA nonfarm payrolls.
- Update the monthly start/end date.
start_f
should be the start of the forecast period (generally the current month).end_f
must be 11 months afterstart_f
. - Make sure the working directory is pointing at the right folder.
- Update
data.csv
using the latest data from BLS. Make sure to extend the empty rows too. Source (NSA Payrolls): https://data.bls.gov/timeseries/CEU0000000001 Source (SA Payrolls): https://data.bls.gov/timeseries/CES0000000001 - Run from command line:
python arima_jobs_day.py
Note: First time use may require installation of pyramid.arima, matplotlib and statsmodels Python packages. Requires Python 3.5. Mac Terminal commands:
pip install pyramid-arima
pip install -U statsmodels
pip install matplotlib
See requirements.txt
for version requirements. Newer versions of statsmodels appear to be incompatible with pyramid.arima
. Installing the exact versions required can be accomplished with the following code:
pip install -r requirements.txt
Author: Andrew Chamberlain, Ph.D.
Glassdoor Economic Research (glassdoor.com/research)
October 2018