The United States Department of Homeland Security publishes several reports to provide insight into their repsonses to requests for information under the Freedom of Information Act. This is a quick look at how those numbers convey the current status of their work in this area.
In the executive summary of the 2018 Freedom of Information Act Report to the Attorney General of the United States and the Director of the Office of Government Information Services, it is stated that one of the focus points of their efforts in the 2018 fiscal year was to reduce their backlog of open requests. I will try to see if this data contains any indication of that being the case.
This graph shows that most of the agencies whose data I examined closed over half of the requests they worked on throughout the 2018 fiscal year. In fact, almost half of the agencies managed to close out around 80% or more of their requests.
The next set of charts is a breakdown of the current status of cases by year they were initially requested for each agency I looked at. Based on the chart above, we should see more ‘Closed’ requests (in red) than ‘Open’ requests (in blue); and these charts validate that assumption. If DHS’s focus on reducing the backlog of requests had any effect, we should see some indication of that in the years preceeding 2018.
I had hoped to see a strong indication of focused effort on closing out backlogged requests. The charts didn’t represent that to me, so I ran a chi-squared hypothesis test to see if a request’s current status had any relationship to being a backlogged request. The results showed a definite dependence between the request status and whether it was a backlogged request, but the data needs further analysis to determine what that dependence is based on.
My analysis did not dive into the final disposition of closed requests. They could have been granted, denied, or denied in part. Denials could have been for several different reasons. If you would like to explore it further, I stored the full spreadsheets I used in this github repository. Alternatively, you could 
and follow along as I reduce this:
| Request # | Agency | Initial Request Type | Track Type | Initial Received Date | Perfected | Case Status | Days Opened | Closed Date | Days to Process | Disposition | Disposition “Other” Reason | Exemptions Used |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2017-OBFO-02364 | CISA | OBIM FOIA | Complex | 2016-11-04 00:00:00 | NaT | Open | nan | NaT | nan | nan | nan | nan |
| 2017-OBFO-03060 | CISA | OBIM FOIA | Complex | 2016-10-21 00:00:00 | NaT | Open | nan | NaT | nan | nan | nan | nan |
| 2017-OBFO-03260 | CISA | OBIM FOIA | Simple | 2016-11-28 00:00:00 | NaT | Closed | nan | 2018-01-29 00:00:00 | nan | Other Reasons | nan | nan |
| 2017-OBFO-03857 | CISA | OBIM FOIA | Complex | 2016-12-07 00:00:00 | NaT | Closed | nan | 2017-10-23 00:00:00 | nan | Granted/Denied in Part | (b)(6),(b)(7)(C),(b)(7)(E),(k)(2) | nan |
| 2017-OBFO-02413 | CISA | OBIM FOIA | Complex | 2016-11-03 00:00:00 | NaT | Open | nan | NaT | nan | nan | nan | nan |
to this:
| Agency | Case Status | Fiscal Year Requested | Backlogged Request |
|---|---|---|---|
| CISA | Open | 2017 | True |
| CISA | Open | 2017 | True |
| CISA | Closed | 2017 | True |
| CISA | Closed | 2017 | True |
| CISA | Open | 2017 | True |
in order to easily plot the data I was interested in.
This also allowed me to generate this contingency table:
| Case Status | New This Year | Backlogged |
|---|---|---|
| Closed | 107891 | 5737 |
| On Hold | 575 | 10 |
| Open | 8217 | 2713 |
to use in the chi-squared tests which resulted in chi-statistic of 6208.212370170039 and a p-value of 0.0.
Let me know if you come up with anything interesting!