the more i dig, the more i find that 'generic duns' tied to multiple vendors, multiple addresses.
i don't want to get hung up on that one building with that one suite #911 - but damn if it isn't suspicious.
(i guess i'm just saying more to come)
it looks like they are truly several different "designations" within the generic 123456787 DUN - but i found one 'zip' where every order is tied to that suite 911, and the "place of performance" is always SA.
and then there are these. pretty big orders in 2003 for "other buildings" in afghanistan.
what's the point of a generic DUN?
is there a legitimate reason?
this search, then, seems excessively large.
trying to figure out a way to narrow it down, or at least another way to categorize the different groups of activity.
there are definitely patterns, but 940k contracts is kinda hard to make sense of.
oh, here, this is nice. for that dun, it lists the top ten vendors.
MISCELLANEOUS FOREIGN CONTRACTORS (523070)
MISCELLANEOUS FOREIGN AWARDEES (392444)
MISCELLANEOUS FOREIGN CONTRACT (13203)
J & B TRUCK REPAIR SERVICE (2176)
UNICOR (60)
FROESCH GMBH & CO.KG (56)
INTERDEAN AG (46)
ANTEON CORPORATION (39)
WEST COAST WELDERS (36)
ELTON OIL COMPANY (35)
specific dates? tell me what you're looking for, i'm archiving now.
their system hard-stops at 30,000 lines, so i have to pull down everything in batches.
i'm into 2006, and i'm going to keep on grabbing before i try to sort or nail down.
tell me expressly what you're looking for, and i'll find it.
that's what i do.
i have all of the data, importing it in quarterly increments.
will analyze and spit out something when i'm done.
it'll be a while, i have a few other pressing issues today…but i have the data offline, and i will be back today.
hey - that's not what i said.
there are multiple place of performance for the DUNS, it's just that there are some specific reference numbers which always bring up SA.
sorting, will report more later.
https:// www.fpds.gov/ezsearch/fpdsportal
search string
VENDOR_DUNS_NUMBER:"123456787" SIGNED_DATE:[2018/01/01,2018/05/31]
the .csv downloader will only spit out 30,000 lines at a time, and there are 940,000 records from 2000 - 2018, so you need to parse it appropriately. i went by years, then by six month periods, and for some time spans i could only get two months at a time.
d/l
import
reassemble
analyze
report
good to have multiple eyes on it - i'm on step four.