Auditing DOGE's Receipts
What is DOGE?
The Department of Government Efficiency was organized by an executive order which renamed the previous United States Digital Service to "DOGE".
This Executive Order establishes the Department of Government Efficiency to implement the President's DOGE Agenda, by modernizing Federal technology and software to maximize governmental efficiency and productivity.
Before continuing, visit the DOGE site to see it for yourself.
When landing on the site, the first thing you see is a wall of tweets with the header "Latest Work". All of these posts link to X.
One thing to note is that in the logo text there is a slogan of "The people voted for major reform". There are six navigation links in the header that lead to savings, spend, workforce, regulations, join, and about.
Savings
At the top of the "Savings" page it says DOGE is estimating a total of $65B in savings.
Below, there are two tables of interest. One table lists contracts, and the other leases.
The tables lack summaries and appear to be in a random order. I thought it would be nice to analyze the data to better understand what these tables are really presenting.
I scraped the data and crunched some numbers which are presented below.
Aggregate Values as of 2/24/2025
| High Level Data | Value |
|---|---|
| Number of Contracts | 1108 |
| Number of Leases | 97 |
| TOTAL Column Values | Value |
|---|---|
| Contract Sum | 12,673,048,099.0 |
| Leases Sum | 78,852,363.0 |
| Total (Contracts + Leases) | 12,751,900,462.0 |
| % of Estimated Savings | 19.62 |
| Contract % of Estimated Savings | 19.5 |
| Leases % of Estimated Savings | 0.12 |
| SAVINGS Column Values | Value |
|---|---|
| Contract Sum | 7,187,588,305.0 |
| Leases Sum | 144,574,545.0 |
| Total (Contracts + Leases) | 7,332,162,850.0 |
| % of Estimated Savings | 11.28 |
| Contract % of Estimated Savings | 11.06 |
| Leases % of Estimated Savings | 0.22 |
10 Largest Contracts (Saved Value)
| Agency | Link | Saved | % Of Saved |
|---|---|---|---|
| USAID | Link | 654,990,000 | 9.11 |
| SSA | Link | 231,864,794 | 3.23 |
| USAID | Link | 167,430,251 | 2.33 |
| USAID | Link | 159,521,884 | 2.22 |
| USAID | Link | 143,948,652 | 2.00 |
| USAID | Link | 132,770,259 | 1.85 |
| USAID | Link | 114,894,279 | 1.60 |
| DOT | Link | 94,000,000 | 1.31 |
| USAID | Link | 92,024,980 | 1.28 |
| USAID | Link | 79,970,524 | 1.11 |
| Aggregate | N/A | 1,871,415,623 | 26.04 |
In Summary
$57.7B of the estimated $65B in savings (89%) is unaccounted for in these receipts. It is unclear if the savings are just pure contract values or annual savings. It is also assumed that the savings values are correct in the table data itself.
Technical Notes OPTIONAL
While writing this article, DOGE actually changed the structure of the website and took down their /api and started server side rendering. This made analyzing the data a tiny bit more complicated. Below are my notes on these changes.
Site Architecture before 2/24
The old version of the site would fetch data from /api/receipts/overview. The data from this endpoint looked like this at the top level:
{
contracts: [
{
"date": "1/31/2025",
"piid": "1232SA22P0086",
"agency": "Department of Agriculture",
"ceiling_value": 70000,
"value": 0,
"update_date": "2/6/2025",
"fpds_status": "CLOSE OUT",
"fpds_link": "https://www.fpds.gov/ezsearch/jsp/viewLinkController.jsp?agencyID=12H2&PIID=1232SA22P0086&modNumber=P00002&idvAgencyID=&idvPIID=&contractType=AWARD",
"vendor": "STRATEGIC DIVERSITY INITIATIVES INC",
"description": "DIVERSITY AND INCLUSION CONSULTING AND TRAINING, TO INITIATE AND COMPLETE A 3 YEAR STRATEGIC PLAN, FACILITATE MEETINGS, LEAD TRAINING AND REVIEW DIGITAL CONTENT."
},
...
],
leases: [
{
"sq_ft": "3,076",
"date": "1/22/25",
"location": "Miami, FL",
"description": "Lease end date 2/28/25.",
"ceiling_value": "112278",
"value": "7383",
"agency": "Department of State"
},
...
]
}Site Architecture after 2/24
Now, the data is embedded in HTML body script tags. You can parse this data two ways - 1) parse the scripts and re-assemble, 2) parse the rendered HTML (after showing all and toggling columns). The latter is much easier initially, but will not contain as much data.
An example of the data embedded within script tags looks like:
<script>self.__next_f.push([1," Protection Bureau\",\"ceiling_value\":\"305382.00\",\"value\":\"0.00\",\"update_date\":\"2/14/2025\",\"fpds_status\":\"EXERCISE AN OPTION\",\"fpds_link\":\"https://www.fpds.gov/ezsearch/jsp/viewLinkController.jsp?agencyID=955F\u0026PIID=9531CB20F0053\u0026modNumber=P00006\u0026idvAgencyID=\u0026idvPIID=NNG15SC05B\u0026contractType=AWARD\",\"vendor\":\"COUNTERTRADE PRODUCTS, INC.\",\"description\":\"LEAPTREE OPTIMIZE LICENSES\"},{\"date\":\"2/12/2025\",\"piid\":\"20343024F00036\",\"agency\":\"Consumer Financial Protection Bureau\",\"ceiling_value\":\"97492.00\",\"value\":\"0.00\",\"update_date\":\"2/14/2025\",\"fpds_status\":\"BASE AWARD\",\"fpds_link\":\"https://www.fpds.gov/ezsearch/jsp/viewLinkController.jsp?agencyID=2036\u0026PIID=20343024F00036\u0026modNumber=0\u0026idvAgencyID=\u0026idvPIID=NNG15SC23B\u0026contractType=AWARD\",\"vendor\":\"ACCESSAGILITY LLC\",\"description\":\"LIGHTBEND AKKA ENTERPRISE SUBSCRIPTION\"},{\"date\":\"2/12/2025\",\"piid\":\"9531CB24F0037\",\"agency\":\"Consumer Financial Protection Bureau\",\"ceiling_value\":\"441306.00\",\"value\":\"0.00\",\"update_date\":\"2/14/2025\",\"fpds_status\":\"BASE AWARD\",\"fpds_link\":\"https://www.fpds.gov/ezsearch/jsp/viewLinkController.jsp?agencyID=955F\u0026PIID=9531CB24F0037\u0026modNumber=0\u0026idvAgencyID=\u0026idvPIID=NNG15SC23B\u0026contractType=AWARD\",\"vendor\":\"ACCESSAGILITY LLC\",\"description\":\"ENTERPRISE DB POSTGRES ADVANCED SERVER RENEWAL\"},{\"date\":\"2/12/2025\",\"piid\":\"20343024F00012\",\"agency\":\"Consumer Financial Protection Bureau\",\"ceiling_value\":\"16309.00\",\"value\":\"0.00\",\"update_date\":\"2/14/2025\",\"fpds_status\":\"BASE AWARD\",\"fpds_link\":\"https://www.fpds.gov/ezsearch/jsp/viewLinkController.jsp?agencyID=2036\u0026PIID=20343024F00012\u0026modNumber=0\u0026idvAgencyID=\u0026idvPIID=NNG15SC70B\u0026contractType=AWARD\",\"vendor\":\"DH TECHNOLOGIES, LLC\",\"description\":\"APPLECARE OS SUPPORT - PREFERRED\"},{\"date\":\"2/12/2025\",\"piid\":\"20343021F00029\",\"agency\":\"Consumer Financial Protection Bureau\",\"ceiling_value\":\"2038330.00\",\"value\":\"0.00\",\"update_date\":\"2/14/2025\",\"fpds_status\":\"EXERCISE AN OPTION\",\"fpds_link\":\"https://www.fpds.gov/ezsearch/jsp/viewLinkController.jsp?agencyID=2036\u0026PIID=20343021F00029\u0026modNumber=P00006\u0026idvAgencyID=\u0026idvPIID=NNG15SC71B\u0026contractType=AWARD\",\"vendor\":\"FCN, INC.\",\"description\":\"ANOMALI LICENSES FOR CFPB\"},{\"date\":\"2/12/2025\",\"piid\":\"20343021F00075\",\"agency\":\"Consumer Financial Protection Bureau\",\"ceiling_value\":\"1024613.00\",\"value\":\"0.00\",\"update_date\":\"2/14/2025\",\"fpds_status\":\"EXERCISE AN OPTION\",\"fpds_link\":\"https://www.fpds.gov/ezsearch/jsp/viewLinkController.jsp?agencyID=2036\u0026PIID=20343021F00075\u0026modNumber=P00004\u0026idvAgencyID=\u0026idvPIID=NNG15SC71B\u0026contractType=AWARD\",\"vendor\":\"FCN, INC.\",\"description\":\"SURVEY TOOL (QUALTRICS)\"},{\"date\":\"2/12/2025\",\"piid\":\"20343024F00028\",\"agency\":\"Consumer Financial Protection Bureau\",\"ceiling_value\":\"152587.00\",\"value\":\"0.00\",\"update_date\":\"2/14/2025\",\"fpds_status\":\"BASE AWARD\",\"fpds_link\":\"https://www.fpds.gov/ezsearch/jsp/viewLinkController.jsp?agencyID=2036\u0026PIID=20343024F00028\u0026modNumber=0\u0026idvAgencyID=\u0026idvPIID=NNG15SC71B\u0026contractType=AWARD\",\"vendor\":\"FCN, INC.\",\"description\":\"SOLARWINDS MAINTENANCE AND SUPPORT\"},{\"date\":\"2/12/2025\",\"piid\":\"9531CB22F0104\",\"agency\":\"Consumer Financial Protection Bureau\",\"ceiling_value\":\"681980.00\",\"value\":\"0.00\",\"update_date\":\"2/14/2025\",\"fpds_status\":\"CHANGE ORDER\",\"fpds_link\":\"https://www.fpds.gov/ezsearch/jsp/viewLinkController.jsp?agencyID=955F\u0026PIID=9531CB22F0104\u0026modNumber=P00003\u0026idvAgencyID=\u0026idvPIID=NNG15SC71B\u0026contractType=AWARD\",\"vendor\":\"FCN, INC.\",\"description\":\"DATABRICKS ENTERPRISE PLATFORM, DATABRICKS PLATFORM CAPACITY, AND ADDITIONAL SUPPORT.\"},{\"date\":\"2/12/2025\",\"piid\":\"9531CB23F0039\",\"agency\":\"Consumer Financial Protection Bureau\",\"ceiling_value\":\"263402.00\",\"value\":\"0.00\",\"update_date\":\"2/14/2025\",\"fpds_status\":\"BASE AWARD\",\"fpds_link\":\"https://www.fpds.gov/ezsearch/jsp/viewLinkController.jsp?agencyID=955F\u0026PIID=9531CB23F0039\u0026modNumber=0\u0026idvAgencyID=\u0026idvPIID=NNG15SC71B\u0026contractType=AWARD\",\"vendor\":\"FCN, INC.\",\"description\":\"NUTANIX HARDWARE DESKTOP ENGINEERING LAB SYSTEM UPGRADE\"},{\"date\":\"2/12/2025\",\"piid\":\"9531CB24F0075\",\"agency\":\"Consumer Financial Protection Bureau\",\"ceiling_value\":\"446087.00\",\"value\":\"0.00\",\"update_date\":\"2/14/2025\",\"fpds"])</script>This data is contained in __next_f variable which can then be assembled into the data structure returned from the deprecated api above. However, now its embedded server-side.
Getting The Data
The simplest way to get the data is to just parse the HTML tables and convert them into structures that make it easy to analyze the data. I used Beautiful Soup and Pandas libraries in Python to extract and analyze the data.