USING PDQ EXPLORE FOR PUMS 5% DATA
PDQ Explore is a high-speed computer program developed by
Public Data Queries, Inc. and used with permission by the Population
Studies Center for manipulating 1990 Census microdata. It includes both
the PUMS 1% and 5% files.
Instructions for accessing the program are available on the
web through: http://www.pdq.com/pdq_explore.txt. Although UMich
affiliates can access it through their login accounts, those outside the
University may telnet to pdq.psc.lsa.umich.edu (login:demo;
password:demo).
A 24-page listing of Michigan PUMA geography appears at:
http://www.lib.umich.edu/govdocs/pumasmi.html
Below is a very simple search using the 5% PUMS file. The
question: number of people born in China who immigrated to the U.S.
between 1980 and 1980. (In the search below, pob=207 means place of birth
is China; immigr=1..4 means immigration between 1980 and 1990l; state=26
means Michigan; puma=3100..4400 are the public use microdata areas in
southeast Michigan.)
1. Access login.itd.umich.edu [Sign into login account]
2. At the % sign, TYPE
~pops/bin/explore [Brings up Explore from
IFS space]
pdq_explore:
3. TYPE
load pums_1990_5pct [Loads the 5% PUMS
nationally]
pdq_explore:
4. TYPE
select immigr:1..4 [Universe limited to people immigrating
Query type tabulate: 1980-90, eg. 1 or 2 or 3 or 4. A
colon must be used when giving a
range.]
------------------------------------------------------------------------------
Select- 0: immigr:1..4
------------------------------------------------------------------------------
------------------------------------------------------------------------------
query tabulate on dataset pums_1990_5pct at server query.pdq.com:6968.
Table has 1 elements.
==============================================================================
pdq_explore:
5. TYPE [Universe limited to people born in
select pob=207 China. China, Taiwan and Hong
Kong would be pob=207|209|238. The
line means OR. An = sign can be used
for selecting one or OR-ed variables
but not a range]
Query type tabulate:
------------------------------------------------------------------------------
Select- 0: pob=207
Select- 1: immigr:1..4
------------------------------------------------------------------------------
------------------------------------------------------------------------------
query tabulate on dataset pums_1990_5pct at server query.pdq.com:6968.
Table has 1 elements.
==============================================================================
pdq_explore:
6. TYPE
select state=26 [This ensures Michigan]
Query type tabulate:
------------------------------------------------------------------------------
Select- 0: state=26
Select- 1: pob=207
Select- 2: immigr:1..4
------------------------------------------------------------------------------
------------------------------------------------------------------------------
query tabulate on dataset pums_1990_5pct at server query.pdq.com:6968.
Table has 1 elements.
==============================================================================
pdq_explore:
7. TYPE [Row indicates a range of PUMAS
row puma:3100..4400 in Southeast Michigan. Impossible
to identify scattered PUMAS as a
row. Must recode. See recoded
search at end of this example.
To break this down by sex, type
column sex]
Query type tabulate:
------------------------------------------------------------------------------
Select- 0: state=26
Select- 1: pob=207
Select- 2: immigr:1..4
------------------------------------------------------------------------------
Row [ 3100, 4400] : puma:3100..4400
------------------------------------------------------------------------------
query tabulate on dataset pums_1990_5pct at server query.pdq.com:6968.
Table has 1301 elements.
==============================================================================
pdq_explore:
8. TYPE [Add population weight]
weight pwgt1
Query type tabulate:
------------------------------------------------------------------------------
Select- 0: state=26
Select- 1: pob=207
Select- 2: immigr:1..4
------------------------------------------------------------------------------
Row [ 3100, 4400] : puma:3100..4400
Weight: pwgt1
------------------------------------------------------------------------------
query tabulate on dataset pums_1990_5pct at server query.pdq.com:6968.
Table has 1301 elements.
==============================================================================
pdq_explore:
9. TYPE [Execute program]
run
query took 11.422574 seconds to run (including communications overhead and compile time).
Table of frequency counts
N: 12501046 Selected: 144
==============================================================================
<<<<<< 1470
3100 392 Ann Arbor
3196 0
3197 0
3198 0
3199 0
3200 273
3201 0
3202 0
3301 231
3302 0
3303 0
3400 0
3401 0
3402 19
3403 119
3404 23 [Only selected portions of
3405 0 printout]
3406 0
3599 0
3600 38
3601 0
3602 0
3798 0
3799 0
3800 49
3801 0
3901 0
3902 91
3903 0
4000 183
4001 0
4002 0
4102 133
4103 281
4104 256
4105 87
4106 45
4107 95
4108 0
4399 0
4400 0
>>>>>> 0
Total 3785
pdq_explore:
10. TYPE
quit
galaxian%
11.TYPE
logout
RECODED SEARCH
The following search recodes the pumas, so just Ann Arbor,
the remainder of Washtenaw County, and the City of Detroit are
selected. Note that the range of pumas must also be selected in
the universe.
1. Access login.itd.umich.edu
2. % sign: ~pops/bin/explore
3. pdq_explore: load pums_1990_5pct
4. pdq_explore: select immigr:1..4
5. pdq_explore: select pob=207
6. pdq_explore: select state=26
7. pdq_explore: select puma:3100..3308 [New element for recode]
8. pdq_explore: recode new_puma
At this point the computer loads a new program.
Delete the introductory information (CNTRL-K)
and also the #s. Write over the template.
R: new_puma
r: 3100 3100 1 AnnArb
r: 3200 3200 2 Washtenaw
r: 3301 3308 3 Detroit
CNTRL-X to save. You will return to the main
Explore menu.
9. pdq_explore: row new_puma(puma)
10. pdq_explore: weight pwgt1
------------------------------------------------------------------------------
11, pdq_explore: run
<<<<<< 0
AnnArb 392
Washtenaw 273
Detroit 231
>>>>>> 0
Total 896
12. pdq_explore: quit
13. tempest% logout
Alternative Recode
R: new_puma
r: 0 3099 0 Other
r: 3100 3100 1 AnnArb
r: 3200 3200 2 Washtenaw
r: 3301 3308 3 Detroit
r: 3309 9999 0
FUNCTIONS
"Recode" groups together existing values.
"Function" is an algebraic formula, which I don't entirely understand.
EXAMPLE of Function:
F: pums_1990_5pct.housing.hus_race
f: sum ((sex==0 && relat1<=1)*race):0..327
F: pums_1990_5pct.housing.wif_race
f: sum ((sex==1 && relat1<=1)*race):0..327