The corpus's temporal distribution is heavily weighted toward 2015–2020:
| Year | Documents | Notes |
|---|---|---|
| 1995–2005 | 1–5/year | Extremely sparse |
| 2006 | 5 | First plea deal |
| 2008–2009 | 6 each | Incarceration period |
| 2015 | 13 | Virginia Giuffre civil suit |
| 2016 | 101 | Significant increase |
| 2017 | 72 | Sustained activity |
| 2018 | 100 | Pre-arrest peak |
| 2019 | 145 | Arrest year — highest |
| 2020 | 114 | Post-death filings |
2001 has zero dated documents. This is not anomalous — no year before 2006 has more than 5 dated documents. The corpus's date coverage is prosecution-driven, concentrated on the 2015–2020 period when legal proceedings were most active.
| Year | Sources | Notes |
|---|---|---|
| 2001 | 0 datasets | No date-indexed documents |
| 2019 | sec-edgar (96), indexofepstein (49) | 145 total |
| 2020 | sec-edgar (102), indexofepstein (12) | 114 total |
Zero GDELT events mentioning "EPSTEIN" exist for the 2001 or 2019–2020 windows. The GDELT data in the graph may be limited in scope or temporal coverage.