[acs-r] Different data via API vs FactFinder?

[Harald Kliems]

For our bike advocacy organization, I have been looking into commuting
data from the ACS. Until recently, I used the FactFinder to download
tables and then manipulate and analyze the data in LibreOffice Calc.
I've now tried using the very useful ACS.r package to improve my work
flow, but looking at the retrieved data, there seem to be differences
between what I get through the ACS.r/API and what I download through
the web interface from FactFinder.

For example, for 2016 ACS 1-year data from table B08006 for Madison (WI):

###########
## set geography to city of Madison, Wisconsin
Madison <- geo.make(state="WI", place = 48000)

## select only the total, car/truck/van, public transport, bicycle,
walk variables from table B08006
TransitVars <- acs.lookup(2016, span = 1, table.number = "B08006",
dataset = "acs")[c(1,2,8,14,15)]

#fetch data for 2016 year and return them into acs object

ACSData2016 <- acs.fetch(2016, geography = Madison,
                     variable = TransitVars,
                     col.names = c("Total", "Motor Vehicle", "Public
transit", "Bicycle", "Walked"))
##########

Just looking at the "total" column, I get 140,435 through ACS.r but
146715 when looking up the same table and geography through
FactFinder. Similar differences exist for other years and the other
variables.

I'm pretty new to R, and so it's entirely possible that I'm doing
something wrong, but I don't see an obvious explanation for the
discrepancy.

Any hints would be appreciated.

Thanks,
 Harald.