Our updated database shows the 9,771 times the U.S. trained foreign military units between Oct 2019 and Sept 2020

Photo by Meriç Dağlı on Unsplash (adapted by SFM)

Spring has sprung and the U.S. Department of State has just published the Foreign Military Training and DOD Engagement Activities of Interest report for the 2020-2021 fiscal year. Published annually since 2000, these important reports show how the U.S. has spent a large amount of its training and assistance budget in the previous fiscal year, and how they intend to spend it in the present one.

As usual they have released this report as a set of PDFs; also as usual we have liberated the data contained Section IV – Part IV of this report and turned it into data that is machine readable and easier to examine in spreadsheets and other data analysis tools.

See the FY 2020-2021 U.S. Foreign Military Training data here.

This update adds 9,771 new training interventions to our database, which now contains a total of 237,603 trainings covering the two decades from 2000 to September 2020.

You can browse around the data using the various filters and buttons provided by the database, but to get the most of out it you’ll need to use a bit of Structured Query Language (SQL). In most cases, though, you can get the results you need without SQL or by fiddling with and adapting an existing query. Here’s a few SQL statements to get you started:

In addition, you can download the results of your query or all the raw data to use in Excel or other data analysis tools.

Tell us more about how these data have been used

We’re pleased that the data we’ve published has has been used in some interesting ways, for example:

When using the the data, bear in mind they are automatically extracted from the PDFs, cleaned up a bit, and augmented with some metadata. Each individual training has a permenant unique ID number, internationally-standardised start and end dates, and can be traced directly to its source PDF. It also means there may be some surprises in the data – errors that exist in the original publication will come across verbatim; there may also – rarely, in our experience so far – be some errors that are a result of the automated processing.

We have published the technical toolkit we use to do this, which also contains a full description of the fields you’ll find the data. Let us know if you spot errors and other problems that we need to fix.

Resources mentioned in this blog post

%d bloggers like this: