SQL Queries of the Zoning Database
Joseph Ferreira, Jr.

The 'toy' parcel database helped us understand basic SQL queries and learn how to join several small tables with at most a few dozen rows. These tables were similar in structure to tables used to record land use and ownership but contained hypothetical data. We used these small tables to get comfortable with relational thinking and SQL queries that handle one-to-many relationships among a few tables with dozens of rows. Now, we'll focus on exploring, summarizing, and cross-referencing one larger table (of zoning variance information) that has several dozen columns, several thousand rows, and various lookup tables that explain certain codes, locate the zoning variance cases, and relate them to the census demographics of their neighborhood. Later, we'll relate this database to even larger tables (the parcel database for Boston) as well as to maps.

Other class web pages (within the SQL Notes section) describe zoning variance database (and associated lookup tables). This webpage contains sample SQL queries (formatted for Oracle SQL*Plus but generally consistent with MS-Access SQL as well) that explore the 'zoning' table data. The zoning database and related lookup tables are also included within an MS-Access database that is available from the class data locker: http://mit.edu/11.521/data/11.521_lab3.mdb.

The basic steps that we discussed to familiarize ourselves with the data were:

1. Begin with simple SQL selects. Learn some formatting commands to handle columns that are too wide.
2. Use 'group by' to summarize data within groupings and take advantage of SQL's structure and your cut and paste tools to add/modify columns, the 'where' clause and the 'group by's to quickly redo the queries for selected sub-groups.
3. Dump the data into a spreadsheet to allow graphing and summary stats (You can spool the SQL output to a file, edit it slightly, and reading it into Excel or another spreadsheet. Alternatively, you can link Excel to the MS-Access database containing the zoning variance data.)
4. Improve the readability of some of the listings by using the lookup tables to substitute neighborhood names, landuses, etc. for various codes.
5. Develop new lookup tables that help categorize various cases using your own reasoning and discoveries. (We'll be doing more of this in a related lab exercise.)

     set linesize 132
     set pause on
     set pause 'Ready to Rip...'

     column applicant format A10
     column far format 990
     column fname format A10
     column streetname format A10

     select * from cat;

The SQL queries listed below are formatted for Oracle, but they will also work in the MS-Access database, 11.521_lab3.mdb, with minimal changes. You may need to make minor adjustments to the SQL to conform to the MS-Access syntax. For example, you will need to insert the 'as' keyword to rename a column: for example, "select count(*) AS rowcount from...". You may also need to examine the SQL statement in an editor before you paste it into MS-Access to be sure that it does not contain hidden characters (such as &nbsp or extra line feeds). MS-Access may also complain if there are leading blanks in SQL statement lines that have been formatted for easy reading.

Sample Zoning Database Queries

The next few queries just list selected variables. They'll be useful to cut-and-paste variable names into your own queries and to eyeball the data and experiment with formatting columns:

Query 1:

select CASENUMBER,HRINGDATE,APPLICANT,FNAME,STRTNUMBER,
       STRTPREFIX,STREETNAME,STREETTYPE,
       WARD,PRECINCT
  from ZONING;

select CASENUMBER, HRINGDATE, NEIGHBRHD, SUBNGHBRHD,
       CENSUSTRAC,CENSUSBLK,PARCELNUMB,
       EXISTZONIN,ESTCOST,EXISTUSE,PRPSEDUSE,CHNGEOCCUP
  from ZONING;

select CASENUMBER,CHNGEUSE,ADDITION,ERECTSTRUC,ERECTAUXIL,
       LEGALIZE,SUBDIVISIO,PARKING,MISCPURPOS,EXISTUNITS
  from ZONING;

select CASENUMBER,PRPSDUNITS,EXISTSQFT,PRPSDSQFT,
       USEITEM1,UI1,USEITEM2,UI2,USEITEM3,UI3
  from ZONING;

select CASENUMBER, VARIANCE74,
       NONCONF91,NONCONF92,OFFPKG101,LOTSZE141,LOTSZE142,
       LOTSZE143,LOTSZE144,FAR151,HEIGHT161,ROOF168,OSPACE171
  from ZONING;

select CASENUMBER, 
       FYARD181,SYARD191,RYARD201,SETBACK211,OFFPKG231,
       OFFLOAD241,IPOD,
       DESIGNREV,DSCOMPLETD,BRARECOM,BOARDDECIS
  from ZONING;

The above queries use default widths and column formats. Try this one after running the 'column' formatting commands listed above in the previous section.

Query 2:

select CASENUMBER,HRINGDATE,substr(APPLICANT,1,10) AS applicant,
       substr(FNAME,1,10) fname, STRTNUMBER,
       STRTPREFIX,substr(STREETNAME,1,10) AS streetname, STREETTYPE,
       WARD,PRECINCT
  from ZONING;

The property description variables are encoded as Yes=2, No=1 and Unknown=-1.  Hence, the following queries will compute the percentage of non-missing cases that met each condition:

Query 3:

select count(*), 100*avg(CHNGEUSE-1), 100*avg(ADDITION-1),
       100*avg(ERECTSTRUC-1)
  from zoning
 where CHNGEUSE > 0 and ADDITION > 0 and ERECTSTRUC > 0;

select count(*), 100*avg(ERECTAUXIL-1), 100*avg(LEGALIZE-1),
       100*avg(SUBDIVISIO-1)
  from zoning
 where ERECTAUXIL > 0 and LEGALIZE > 0 and SUBDIVISIO > 0;

select count(*), 100*avg(PARKING-1), 100*avg(MISCPURPOS-1)
  from zoning
 where PARKING > 0 and MISCPURPOS > 0;

Most of the possible zoning code violations are encoded similarly.  Here is another query to compute percentages for some Yes/No encoded variables:  (Note, e.g., that SYARD181 indicates whether the side yard regulations in chapter 181 are potentially violated and the 'where' clause forces consideration of only those rows with no missing values for all three conditions)

Query 4:

select count(*), 100*avg(FYARD181 - 1) AS pct_fyard181,
       100*avg(SYARD191-1), 100*avg(RYARD201 - 1), 
       100*avg(SETBACK211 - 1) 
  from ZONING 
 where FYARD181 > 0 and SYARD191 > 0 and RYARD201 > 0 and SETBACK211 > 0;

Here's the SQL query that could be used (with wide/long linesize/pagesize settings) to dump most of the table into a large, flat ASCII file for input into a spreadsheet.

select CASENUMBER,HRINGDATE,APPLICANT,FNAME,STRTNUMBER,
       STRTPREFIX,STREETNAME,STREETTYPE,
       WARD,PRECINCT,
       HRINGDATE, NEIGHBRHD, SUBNGHBRHD,
       CENSUSTRAC,CENSUSBLK,PARCELNUMB,
       EXISTZONIN,ESTCOST,EXISTUSE,PRPSEDUSE,CHNGEOCCUP,
       CHNGEUSE,ADDITION,ERECTSTRUC,ERECTAUXIL,
       LEGALIZE,SUBDIVISIO,PARKING,MISCPURPOS,EXISTUNITS,
       PRPSDUNITS,EXISTSQFT,PRPSDSQFT,
       USEITEM1,UI1,USEITEM2,UI2,USEITEM3,UI3
       VARIANCE74,
       NONCONF91,NONCONF92,OFFPKG101,LOTSZE141,LOTSZE142,
       LOTSZE143,LOTSZE144,FAR151,HEIGHT161,ROOF168,OSPACE171,
       FYARD181,SYARD191,RYARD201,SETBACK211,OFFPKG231,
       OFFLOAD241,IPOD,OTHERVIOLA,
       DESIGNREV,DSCOMPLETD,BRARECOM,BOARDDECIS
  from ZONING;

Here's a shorter SQL query to prepare an ASCII file that could be loaded into a spreadsheet.

Query 5:

select CASENUMBER,NONCONF91,NONCONF92,OFFPKG101,
       LOTSZE141,LOTSZE142,FAR151,
       HEIGHT161,ROOF168,OSPACE171,
       SETBACK211,OFFPKG231,IPOD
  from zoning;

Here are some useful 'group by' statements:

Query 6:

select existuse, count(*) from zoning
 group by existuse;

  EXISTUSE   COUNT(*)
---------- ----------
        -1         38
         0          2
         1        923
         2         27
         3         59
         4        104
         5         65
         6          1
         7         15
         8        249
         9         38
        10        268
        11         12

Query 7:

   select prpseduse, count(*) 
     from zoning
 group by prpseduse;


 PRPSEDUSE   COUNT(*)
---------- ----------
        -1         19
         1       1172
         2         35
         3         45
         4        106
         5         48
         6          2
         7         12
         8        307
         9         51
        10          1
        11          3

select prpseduse, landuse, count(*)as zcases
  from zoning z, use u
 where z.prpseduse = u.use_code
 group by prpseduse, landuse;  

 PRPSEDUSE LANDUSE                     COUNT(*)
---------- ------------------------- ----------
        -1 UNKNOWN USE                       19
         1 HOUSING                         1172
         2 OFFICE                            35
         3 RETAIL                            45
         4 COMMERCIAL                       106
         5 EDUCATION/INSTITUTIONAL           48
         6 ENTERTAINMENT                      2
         7 RECREATION                        12
         8 MIXED USE                        307
         9 PARKING                           51
        10 VACANT LAND                        1
        11 OTHER LAND USE                     3

Note the above table is identical to the previous one except that the text description of each code is added via the 'lookup' in the USE table.

Query 8:

 select existuse,PRPSEDUSE, count(*) from zoning
  where BRARECOM > 0
  group by existuse, PRPSEDUSE
  order by existuse, PRPSEDUSE;

 select chngeoccup, count(*) from zoning
  group by chngeoccup; 
 

CHNGEOCCUP   COUNT(*)
---------- ----------
        -1          4
         1       1042
         2        754
         8          1

Query 9:

 select designrev, count(*) from zoning
  group by designrev; 



 DESIGNREV   COUNT(*)
---------- ----------
        -1        248
         0        664
         1        591
         2        298

Query 10:

 select dscompletd, count(*) from zoning
  group by dscompletd;



DSCOMPLETD   COUNT(*)
---------- ----------
        -1        249
         0        664
         1        695
         2        193

Query 11:

 select brarecom, boarddecis, count(*) from zoning
  where brarecom > 0 and boarddecis > 0
 group by brarecom, boarddecis
 order by brarecom, boarddecis;

  BRARECOM BOARDDECIS   COUNT(*)
---------- ---------- ----------
         1          1        116
         1          2        154
         1          4         13
         1          5         14
         2          1         82
         2          2        587
         2          4         35
         2          5         31
         2          6          7
         4          1          9
         4          2         60
         4          4         29
         4          5         16
         4          6          4
         5          1         14
         5          2         92
         5          4         39
         5          5         99
         5          6          6
         6          1         20
         6          2        113
         6          4         30
         6          5         31
         6          6         13
         7          7         10
        33          1          1
        33          2          7
        33          4          1
        33         33          1

Back to the 11.521 Home Page
Back to the CRON Home Page
Back to the MIT Home Page

Written by Joseph Ferreira, Jr. (7 Oct. 1998).
Edited and converted to HTML by Thomas H. Grayson on 15 Feb. 2000.
Last modified: 14 February 2011 [jf]