Skip to content

Exported data has no data dictionary or other explanation for the meaning of the various records

This inhibits the use of the exported data, as it is unclear what the various data items mean.

For example, in a Facebook export in June 2021, file apps_and_websites_off_of_facebook/your_off-facebook_activity.json contains the following fragment:

{
  "off_facebook_activity_v2": [
    {
      "name": "Albertsons",
      "events": [
        {
          "id": 373927899929418,
          "type": "CUSTOM",
          "timestamp": 1589383920
        },

We can guess that at UNIX Epoch time "1589383920" (March 20, 2020), some event occurred at Facebook partner "Albertsons". We do not know who that partner is (the US grocery chain?), we don't know what type of event occurred ("CUSTOM") and the id is completely incomprehensible as it does not refer to anything else in the export file.

While this is a Facebook example, the same type of problem occurs with many data exports by many companies. (Add your experience with other companies in the comments)

As a good example, Netflix produces a comprehensive document explaining their export

Edited by Johannes Ernst