• California MEIS Process Guide
  • 1 Introduction
  • 2 Software Requirements
    • 2.1 Download Necessary R Libraries
  • 3 Data Requirements
    • 3.1 Employment Data
    • 3.2 USAspending Data
    • 3.3 Additional Data
      • 3.3.1 SmartPay Data via FOIA
      • 3.3.2 Data Obtained Online
      • 3.3.3 Data Provided Raw/Self-Made
  • 4 Setting Up the Repository
    • 4.1 Obtain Repository
    • 4.2 Modify Data Files
  • 5 Data Dictionary for Repository
    • 5.1 MEIS_Methodology Folder
    • 5.2 Data Folder
      • 5.2.1 Raw Folder
      • 5.2.2 Temp Folder
    • 5.3 Output Folder
    • 5.4 SRC Folder
      • 5.4.1 Deprecated Folder
  • 6 Methods
    • 6.1 Reviewing the Parameters File
      • 6.1.1 Sections of the Parameters File
    • 6.2 Reviewing User Specific Files
    • 6.3 Processing the Data
      • 6.3.1 Clearing Environment, and Loading in Packages and Parameters
      • 6.3.2 Loading in Functions
      • 6.3.3 Obtaining the USAspending Data
      • 6.3.4 Filtering the USAspending Data
      • 6.3.5 Error checking the USAspending Data
      • 6.3.6 Manually Fixing Congressional District Errors
      • 6.3.7 Concatenating the USAspending Data
      • 6.3.8 Splitting DOE from USAspending Data
      • 6.3.9 Filtering DOE to National Security-Related Data
      • 6.3.10 Aggregating the Spending Data
      • 6.3.11 Compiling Employment Data
      • 6.3.12 Running For Loops to Generate IMPLAN Activity Sheets
    • 6.4 Using IMPLAN
  • 7 Conclusion/Discussion
  • 8 What’s Next?
  • 9 License
  • References
  • Published with bookdown

California Military Economic Impact Study Process Guide

Section 8 What’s Next?

Looking past this process guide, our team is working to continue improving tasks that make up this project. Moving forward, here are some changes and modifications we hope to implement to improve the process outlined above:

  • Making the entire process for this analysis applicable to multiple states. While our team has worked to do this by thoroughly developing our parameters file, we recognize that certain parts of this process are inevitably California-centric. For example, some of the data files used in this report (such as the dod_county_shares.xlsx file) represent a fix only applicable to California. Our team will consider future process improvements on this point that can make the analysis process even more broadly applicable.
  • Being vigilant of more APIs so that more of the data analysis process can be automated. One of the biggest breakthroughs for this analysis has been developing a code to automate downloading data from USAspending.gov. However, most of the data sources used for this report (such as FedScope or DMDC) do not have clearly defined or available APIs. Our hope is that these websites will continue to develop ways that can make obtaining data easier, and that we will be able to document, code and share that process.
  • Automating the error checking portion of the code. Our current process of manually checking for errors across the USAspending data sources is time and labor intensive. Our team hopes to generate a more streamlined process that harnesses the power of code to lessen the amount of manual time and labor needed to remediate errors from USAspending.
  • Implementing small area estimation in our calculations for county and district apportionment of spending and employment data. Small area estimation provides a more accurate statistical basis for parsing out statewide numbers across smaller local geographies, such as counties and districts. Our team hopes to spend some time learning more about small area estimation in order to more robustly apportion data across counties and districts.
  • Keeping the process up to date with inevitable changes. Of particular concern are future updates to NAICS codes and state’s congressional districts (due in 2022). These changes in code and geography will certainly affect the analysis, and it will be prudent to continue updating and developing our code and documentation to handle these changes.

Ultimately, we hope to develop a one size fits most method for obtaining and analyzing the impacts of federal spending within the United States.