Although it was not an in-depth view and analysis, I did undertake a
quick check of employment here in the Minneapolis-St. Paul area.... more
to see what was coming out of the CTPP files as compared to what we
received from the state Employment Securities offices. Since we utilize
the state figures and NOT the CTPP numbers, the differences have not
been an issue. I can't offer much of an explanation other than to
relate our own findings. In any case they did NOT seem to be random.
Two checks were made. The first one was entirely with the 2000
figures. I mapped the Employment Securities numbers by TAZ and then the
CTPP part 2 numbers by TAZ. In a fairly sizeable number of instances
the two files were substantially different. However, the differences
seem to have been off-set by combining contiguous TAZs. That is to say,
one zone was under-counted by 1500, but the adjacent was over-counted by
1800. Admittedly I simply chalked up the bulk of these as being
"address matching problems". Correctly or not, that was the assumption
I made. Nearly all of these major differences fell into this
The second check was using the 1990 CTPP and 2000 CTPP figures.
Sorting the set "high to low" by differences surprised me at first. The
greatest gain was 13,319 employees; the next 6 TAZs showed gains from
4,000 to 5,600. Upon closer inspection they mostly proved to by TAZs
where major growth actually did occur. The TAZ with the 13,000 gain was
around the airport where our CTPP generated employment numbers always
seemed suspect. We never use them. Why they are so far off the beam I
don't know. Airports seem to be touchy, quirky activity centers.
Those TAZs at the other end of the spectrum (showing major declines)
appeared to be located adjacent to TAZs with major gains. Again, I
chalked it up to address matching. Maybe I'm being lazy, for it could
be something else! In several instances there were TAZ boundary issues
that played into the mix. In a few other TAZs, there actually were
major declines. For example, a rather large publishing company
relocated from the inner city to a suburban location. That move did
cause a big drop in employment that did NOT rebound.
What came out of this review was that in the overwhelming majority of
instances, mapping TAZs with major drops to those with major increases
resulted in a pattern. They seemed to be contiguous to each other. For
those that weren't, there were determining causes that could be
identified. The TAZs were random around the region, but they affected
TAZs with employment and were usually balanced out with adjacent TAZs.
Maybe our region was lucky. You might want to check to see if this
same pattern checks out with adjacent TAZs, if you haven't already done
so. In the case of New York City, I can't imagine there being much of
an address matching problem.
Show replies by date
Hi, everyone -
I am working on a sub-area analysis and ran into some problems. I wonder if
you happen to have similar experiences and would share your
Our travel forecast model covers more than 50 towns in our region and has
The sub-area analysis focuses on an area of 2 towns. The current 25 traffic
analysis zones in these 2 towns will be expanded to 67 zones for a more
sensitive highway alternative analysis. The total number of model zones will
increase from 1174 to 1216.
The sub-area analysis is in the traffic assignment step only. The plan is to
borrow the trip matrices from the first 3 steps of the model and to expand
the trip matrices from 1174*1174 to 1216*1216, so that the matrices can be
used in traffic assignment.
A highway layer with the expanded zone system is available. It has 1216
A network will be created which has the following link attributes:
Link speed and capacity,
Link free flow time, and
Alpha and beta value for the BPR equation.
Traffic assignment will be run from the tool menu using:
The network *.net (1216 zones)
The highway layer (1216 zones)
The trip matrix (1216*1216)
The link traffic volumes generated from this assignment step is very
different compared to the original model link volumes. I wonder if the
additional centroid connectors in the highway network caused this problem.
Your input will be appreciated.
Capitol Region Council of Governments
241 Main Street
Hartford, CT 06106
Phone: 860-522-2217 * 21