Evaluating access to jobs via transit from disparate neighborhoods CliffordKaplan‐GISforWaterResources2014,TermProject
Manor Corridor
Burnet Corridor
Riverside Corridor
Esri, HERE, DeLorme, MapmyIndia, © OpenStreetMap contributors, andthe GIS user community
Created by: Cliff Kaplan | Fall 2014Data Sources: BBC Consulting; Investor Incentives;
Travis County Appraisal District; Housing WorksNAD 1983 StatePlane Texas Central FIPS 4203 (US Feet)¯ 0 1.5 30.75 Miles
Legend
Number of Units
Less than 50
51 - 100
101 - 200
More than 200
Manor Corridor
Burnet Corridor
Riverside Corridor
Esri, HERE, DeLorme, MapmyIndia, © OpenStreetMap contributors, andthe GIS user community
Created by: Cliff Kaplan | Fall 2014Data Sources: CapMetro
NAD 1983 StatePlane Texas Central FIPS 4203 (US Feet)¯ 0 2.5 51.25 Miles
Legend
CapMetro bus lines
Manor Corridor
Burnet Corridor
Riverside Corridor
Esri, HERE, DeLorme, MapmyIndia, © OpenStreetMap contributors, andthe GIS user community
Created by: Cliff Kaplan | Fall 2014Data Sources: http://.onthemap.ces.census.gov,
U.S. Census BureauNAD 1983 StatePlane Texas Central FIPS 4203 (US Feet)¯ 0 2 41 Miles
Legend
Jobs!( 20 - 300
!( 301 - 1000
!( 1001 - 5000
!( More than 5000
1
Summary
For this term project I set out to design and conduct an easily replicable
geospatialanalyticalmethodtocompareurbancorridors in termsof theaccess to
low‐incomejobsviapublictransitthateachcorridorhas.ThemethodIdeviseduses
nationallyavailablejobsdata,publiclyavailabletransitdatathatisgenerallyeasyto
find incitieswith transit systems,andacollectionofhousingdatasets thatwere
moredifficulttoassemble,butthatwouldbenecessaryforanyonewhoisworking
geospatiallyonaffordablehousing issues tohave.Theanalyticalmethod relieson
simpleGIS skills and is easilymodified according to the level of specificity in the
data(e.g.,numberofbuildingsversusnumberofunits,andwhetherornot transit
headwaydataisavailable).
Conducting the analysis for three urban corridors in Austin, TX—Manor
Road, Burnet Road, and East Riverside Drive—I found that low‐wage jobs (by
definition the relevant class of jobs for people living in affordable housing,) are
twiceasaccessiblefromtheBurnetcorridorastheyarefromtheRiversidecorridor
andnearlyfourtimesasaccessibleastheyarefromtheManorcorridor.Affordable
housing in the Manor corridor has 76% less accessibility to jobs than does the
affordable housing in the Burnet corridor, although the actual number of jobs
reachable from the Manor corridor is only 33% less than the number reachable
from the Burnet corridor. The differences in accessibility are due not only to the
placesthebusesgo,buttowheretheaffordablehousingisinrelationtothenearest
buslinesandhowoftenthebusescome.
2
Contextforthisproject
The Green and Inclusive Corridors project is a U.S. Housing and Urban
Development(HUD)fundedprojectledbyDr.ElizabethMuelleroftheCommunity
andRegionalPlanningdepartmentattheUniversityofTexasatAustin.Theproject’s
objective is to design and conduct a replicable methodology that planners and
advocatescanusetoevaluatethreatstoaffordablehousingfromdevelopmentand
to identify opportunities for protecting affordable housing in their cities. As Dr.
Mueller’sgraduateresearchassistant,Iamhelpingtodesignarubricforcomparing,
acrossmanycriteria,“activitycorridors”indifferentpartsofthecity.Onecriterion
bywhichwewould liketocomparethecorridors is transitaccessto jobs for low‐
incomeresidentsofeachrespectivecorridor.Thistermprojectrepresentsmyeffort
to design a replicable geospatial analytical method for evaluating transit
connectivity fromaffordablehousing to jobs invariouspartsof the city.Although
theGreenandInclusiveCorridorsprojectcompareseightcorridorsinAustin,forthe
purposes of developing the methodology and presenting it in this paper, I have
narrowedthestudyareastothreecorridors.
ProjectObjective
Theobjectiveofthisprojectistodesignandconductareplicablegeospatial
analyticalmethodforcomparingdifferentneighborhoodsintermsofthenumberof
3
low‐paying jobs that are easily accessible by public transit from each
neighborhood.1
Datarequiredfortheanalysis
TheanalysisIhavedesignedrequiresthefollowingdata,regardlessofthecityto
whichtheanalysisisapplied.
Boundarydatafortherespectivestudyareastobecompared.
Locationdata for theaffordablehousingstock. If informationregarding the
number of units in affordable buildings is available, this can be used to
improvetheprecisionoftheanalysis.
Routedataforthepublictransitsystem.Ifheadwaydataisavailable,thiscan
beusedtoimprovetheprecisionoftheanalysis.
Locationdataforrelevantjobs.Asthisanalysisisconcernedwithaffordable
housing it is assumed that the housing residentswork low‐paying jobs, so
onlytheseareincludedintheanalysis.
MeetingthesedatarequirementsintheAustincontext
For this analysis in Austin, I met each of the data requirements in the
followingways.
1Forthepurposesofthisanalysis,wetook“accessibility”tomeanaccessiblewithouthavingtotransferbuses.InAusitn,transferringbetweenbuslinestendstorequirelongwaitingtimeswhichmakesitfarfromidealforcommuters.Althoughpeoplewithoutotheroptionsdocommutethisway,thisanalysisisdesignedtoproduceameasureofqualitysoweareonlyincludingtheone‐buscommutes.
4
Boundarydata for the respective studyareas tobecompared. As previously
noted, the larger project of which this geospatial analysis is a part is a corridor
project that looks at eight corridors in Austin. Dr. Mueller, my colleague Sara
McTarnaghan, and Idecided todrawour studyboundaries ashalf‐milebuffers to
eachsideofthecorridor.SaraMcTarnaghancreatedtheshapefilesweareusingin
thisproject according to this specification. In thispaper, I focuson threeof these
corridors: Burnet Road,Manor Road, and East Riverside Drive. Map 1 shows the
extentandlocationsofthesethreestudycorridors.
Map1:Thethreestudyareas
Manor Corridor
Burnet Corridor
Riverside Corridor
IH 3
5
US
183
LAM
AR
MO
-PA
C
US 290
BEN WHITE
SH 71
CO
NG
RE
SS
FM 2222
CAM
ER
ON
CESAR CHAVEZ
LOO
P 3
60
KOENIG
FM 969MARTIN LUTHER KING
DE
SS
AU
NORTHLANDALLANDALE
US 290
LOO
P 360
Esri, HERE, DeLorme, MapmyIndia, © OpenStreetMap contributors, andthe GIS user community
Created by: Cliff Kaplan & Sara McTarnaghan | Fall 2014Data Sources: City of Austin GIS Data
NAD 1983 StatePlane Texas Central FIPS 4203 (US Feet)¯ 0 1.5 30.75 Miles
5
Locationdatafortheaffordablehousingstock. The datawe use to track the
affordable housing stock in Austin comes from three separate sources, each for a
differenttypeofhousingstock.FromBBCConsulting,weobtainedaddressdataon
government‐subsidized multi‐family affordable housing, which we geocoded to
create a point feature‐class. This data included information about the quantity of
unitsateachbuilding.FromInvestorIncentives,arealestatedatafirm,weobtained
address information for Class Cmulti‐family buildings withmore than 50 units.2
Thisdataincludedinformationaboutthequantityofunitsatdifferentrentalvalues
andsizes,fromwhichwewereabletoselectforinclusionintheanalysisonlythose
buildingsthatmeetcurrentaffordabilitystandardssetbytheHousingAuthorityof
theCityofAustin.3Fordataonnon‐subsidizedbuildingsof less than50units,we
hadtocombineinformationfromtwosourcestocreateanestimate.FromtheTravis
CountyAppraisalDistrict (TCAD),weacquireddata that included theaddressand
numberofunitsforallClassCbuildingsinthecity,fromwhichweusedonlythose
withfewerthan50units.ThisTCADdatadidnotincluderentalvaluesandthusdid
not indicate which buildings were uniformly affordable, which were uniformly
unaffordable to low‐income residents, and which had both affordable and
unaffordableunitsandinwhatproportions.Asurvey(N=50)ofClassCbuildingsof
less than 50 units, conducted last summer by Housing Works, found that
2ClassCbuildings“aretypicallymorethan20yearsoldandlocatedinlessthandesirablelocations.Thepropertyisgenerallyinneedofrenovation,includingupdatingthebuildinginfrastructuretobringituptodate.Asaresult,ClassCbuildingstendtohavethelowestrentalratesinamarketwithotherClassAorClassBproperties”(FromRealtyMogul:https://www.realtymogul.com/blog/what‐does‐it‐mean‐when‐a‐property‐is‐class‐a‐class‐b‐or‐class‐c).3AsofApril1st,2014,theseaffordabilitystandardsare:Efficiency‐$696;1bedroom‐$853;2bedroom‐$1074;3bedroom‐$1454.
6
approximately 25% of units surveyed met affordability standards. Therefore, we
usedtheaddressdataobtainedfromTCADandmultipliedthenumber‐of‐unitsdata
thereinby0.25 toestimate thenumberand locationsofaffordableunits inmulti‐
family buildings of less than 50 units. All of the above data came in the form of
spreadsheets. We geocoded it ourselves to create point feature‐classes. For the
purpose ofmy analysis, itwas not necessary tomaintain the distinction between
thesethreetypesofaffordablehousingstock,soImergedthethreefeature‐classes
into one single shape file containing a point for every building in Austin with
affordable units. Map 2 shows the locations and sizes (number of units) of the
affordablehousingstockwithinthestudyareas.
Routedataforthepublictransitsystem.AlthoughAustinhasa rail lineused
bysomecommuters, thisanalysis focusesexclusivelyonthemuchmoreextensive
bus system. From CapMetro, we acquired a line feature class containing a line
feature foreachpublicbus line.Map3 shows the full extentof thebus system in
Austin,asitrelatestothelocationsofthestudyareas.
Location data for relevant jobs. From the U.S. Census Bureau website,
http://onthemap.ces.census.gov/, Iexportedapointshape file thathasapoint for
everycensusblockinAustin,withafieldthatcontainsthenumberofjobswithinthe
$15,000to$40,000annualpayrange. (Thedata is from2011).Map4showsthe
spatialdistributionofjobswithinthe$15,000to$40,000annualpayrange,asthey
relatetothelocationsofthestudyareas.
Of thedata required for this analysis, thehousing stockdatawas themost
complicatedtoacquireandthiswilllikelybethecaseanywherethisanalysisis
7
conducted.Transitroutedatatendstobeeasytoacquireincitiesthathavetransit
systems.ThejobsdataIusedisavailablefortheentireUnitedStates.
Map2:Thelocationsandsizesoftheaffordablehousingstockwithinthestudyareas
Manor Corridor
Burnet Corridor
Riverside Corridor
Esri, HERE, DeLorme, MapmyIndia, © OpenStreetMap contributors, andthe GIS user community
Created by: Cliff Kaplan | Fall 2014Data Sources: BBC Consulting; Investor Incentives;
Travis County Appraisal District; Housing WorksNAD 1983 StatePlane Texas Central FIPS 4203 (US Feet)¯ 0 1.5 30.75 Miles
Legend
Number of Units
Less than 50
51 - 100
101 - 200
More than 200
8
Map3:TheCapMetrobussysteminAustin,TX
Manor Corridor
Burnet Corridor
Riverside Corridor
Esri, HERE, DeLorme, MapmyIndia, © OpenStreetMap contributors, andthe GIS user community
Created by: Cliff Kaplan | Fall 2014Data Sources: CapMetro
NAD 1983 StatePlane Texas Central FIPS 4203 (US Feet)¯ 0 2.5 51.25 Miles
Legend
CapMetro bus lines
9
Map4:Jobswithinthe$15,000to$40,000annualpayrange
Manor Corridor
Burnet Corridor
Riverside Corridor
Esri, HERE, DeLorme, MapmyIndia, © OpenStreetMap contributors, andthe GIS user community
Created by: Cliff Kaplan | Fall 2014Data Sources: http://.onthemap.ces.census.gov,
U.S. Census BureauNAD 1983 StatePlane Texas Central FIPS 4203 (US Feet)¯ 0 2 41 Miles
Legend
Jobs!( 20 - 300
!( 301 - 1000
!( 1001 - 5000
!( More than 5000
10
Outliningtheobjectivesoftheanalysis
To accurately describe and compare different corridors in terms of the
number of low‐paying jobs that are easily accessible by public transit from
affordablehousingwithineachcorridor,theanalysismustconsider:
Thebuslinesservingeachcorridor,
thenumberofjobstowhichthosebuslinesprovideaccess,
thefrequencywithwhichthebusespassthroughthecorridor,and
theportionofaffordableunitswithaccesstoeachbusline.
Conductingtheanalysis
The analysis that I have designed is conducted on one corridor at a time,
allowingtheanalysttocomparetheresultsamongthecorridorsattheend.
Forclarity,Iwilldescribethestep‐by‐stepprocessofanalysisfortheManor
corridor. In the Results section, Iwill provide the results of analysis for all three
corridors,alongwithillustrativemaps.
Step1:Identifythebuslinesthatservethestudyareaandcreateafeatureclass
ofonlythoselines.
Use Select by location to select from the bus lines layer those lines that
intersectthestudyarea.Then,useExportdatatocreateafeatureclassthatisonly
those linesthatservethestudyarea.UsetheAttributeTable for thisnewlayerto
examinethelinesandtheirrelevancetothisanalysis.InthecaseofAustin,thebus
data does not include information about hours of operation or the bus headways
(which correspond to the number of buses on the line at a given time.) For this
information,IusedbusschedulesfromtheCapMetrowebsiteandImanuallyadded
11
thisinformationtoafieldIcreatedintheAttributeTable,calledBus_at_RH(Buses
atrushhour).Ichosetoeliminatecertainbusesthatappearedtonotbeusefulfor
getting to and fromwork formost people. (For instance, I eliminated UT Shuttle
BuseswhichruntoandfromtheUniversityofTexascampusandareoperatedby
CapMetrofortheexclusiveuseoftheuniversitycommunity.)
Afterdeletingthebuslinesthatarenotrelevanttothisstudy,thereremain
sixlinesthatservetheManorcorridor,showninMap5.
Figure1:TheAttributeTablecontainingthesixbuslinesthatservetheManorCorridorthatI included inthisstudy,withtheaddedfieldforthenumberofbusesthatpassduringrushhour.
Step2:Count the totalnumber of jobs (in the relevantpaybracket) thatare
accessibleonthesebuslines.
UseSelectbylocationtoselectfromthejobslayerthosejobs(pointfeatures)
that are within 0.5 miles of the newly created bus line layer.4Use the Statistics
dialoguebox,accessiblethroughthejobslayerAttributeTabletosumthequantity
ofhighlightedjobs.Makeanoteofthisnumber;itwillbeusedlaterintheanalysis.
Map6showsthejobsthatareaccessiblebybusfromtheManorcorridor.Itisnot
40.5milesisaconventionalestimateforthedistancethatpeoplewillgenerallybewillingtowalktoorfrompublictransportonadailybasis.
12
Map5:ThebusesthatservetheManorcorridor
Esri, HERE, DeLorme, MapmyIndia, © OpenStreetMap contributors, andthe GIS user community
Created by: Cliff Kaplan | Fall 2014Data Sources: CapMatro
NAD 1983 StatePlane Texas Central FIPS 4203 (US Feet)¯ 0 2 41 Miles
Legend
Bus lines serving Manor
Manor Corridor
13
necessarytocreateashapefilecontainingonlythesejobs;theanalysiscanbedone
leaving the original jobs shape file intact. I have exported the highlighted jobs to
create a “Jobs fromManor” point feature class for the purposes ofMap 6, which
showsthejobsaccessiblebytransitfromtheManorcorridor,alongwiththetransit
lines.
Map6:Thelow‐payingjobsaccessiblebytransitfromtheManorcorridor
FromtheManorcorridor,thereare97,151low‐payingjobsaccessiblebybus.
!(
!(!( !( !(
!(
!(!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!( !(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!( !(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!( !(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(!(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(
Esri, HERE, DeLorme, MapmyIndia, © OpenStreetMap contributors, andthe GIS user community
Created by: Cliff Kaplan | Fall 2014Data Sources: CapMatro; U.S. Census Bereau
NAD 1983 StatePlane Texas Central FIPS 4203 (US Feet)¯ 0 2 41 Miles
Legend
Jobs!( 20 - 300
!( 301 - 1000
!( 1001 - 5000
!( More than 5000
Bus lines serving Manor
Manor Corridor
14
Steps3,4,5,arerepeatedsequentially foreachbus line thatserves thestudy
area:
Step3:CountthenumberofjobsaccessiblebybuslineX.
UseSelectbyLocationtoselectthejobswithin0.5milesofbuslineX.Usethe
StatisticsdialogueboxintheAttributeTabletosumthequantityofhighlightedjobs,
andmake a note of this number. In theAttribute Table for the buses serving the
studyarea,createafieldcalledJobsandpopulateitwiththenumberofjobswithin
thereachofeachbusline.Map7(asanexample)showsbusline#18alongwiththe
49,808jobswithin0.5milesofthebusline.Map8(asanotherexample)showsbus
line#350alongwiththe12,412jobswithin0.5milesofthatbusline.
Figure2:Thesameattributetable,nowwiththeadded“Jobs”field.
Step4:Multiplythenumberofbusesatrushhourbythenumberofjobsonthe
buslinetocreateacomposite“Usefulness”score.
The number of jobs on the bus line and the frequency of buses are both
factorsinthe“usefulness”ofthatbuslineforcommuting.AddafieldintheAttribute
TableandusetheFieldCalculator topopulatethe field,multiplyingthenumberof
busesatrushhourbythenumberofjobsreachedbythebusline.
15
Map7:Jobsaccessiblefromthebusline#18
!(
!(!(!( !(
!(!(
!(!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(!(!(
!(
!(
!(!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!( !(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
IH 3
5
LAM
AR
US
18
3
MO
- PA
C
US 290
F
M 2222
CA
ME
RON
CO
NG
RE
SS
CESAR CHAVEZ
KOENIG
MARTIN LUTHER KING
B EN WHITE
FM 969
SH 71
DE
SS
AU
NORTHLAND
Esri, HERE, DeLorme, MapmyIndia, © OpenStreetMap contributors, andthe GIS user community
Created by: Cliff Kaplan | Fall 2014Data Sources: CapMatro; U.S. Census Bereau
NAD 1983 StatePlane Texas Central FIPS 4203 (US Feet)¯ 0 1 20.5 Miles
Legend
Jobs!( 20 - 300
!( 301 - 1000
!( 1001 - 5000
!( More than 5000
Route_18
Manor Corridor
16
Map8:Jobsaccessiblefromthebusline#350
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!( !(
!(
!(
!(
!(
!(
IH 3
5
US
18
3
LAM
AR
MO
-PA
C US 290
SH 71
CO
NG
RE
SS
BEN WHITE
FM
2222
FM973
CA
ME
RON
LOO
P 3
60
FM 969
CESAR CHAVEZ
WILLIAM CANNON
K
OEN IG
MARTIN LUTHER KING
SH
13
0
DE
SS
AU
SLAUGHTER
PEARCE
NORTHLAND
LOO
P 360
US 290
Esri, HERE, DeLorme, MapmyIndia, © OpenStreetMap contributors, andthe GIS user community
Created by: Cliff Kaplan | Fall 2014Data Sources: CapMatro; U.S. Census Bereau
NAD 1983 StatePlane Texas Central FIPS 4203 (US Feet)¯ 0 2 41 Miles
Legend
Jobs!( 20 - 300
!( 301 - 1000
!( 1001 - 5000
!( More than 5000
Bus line #350
Manor Corridor
17
Figure3:The sameattribute table,nowwith theadded “B_x_J” (Busx Jobs)field,whichisanexpressionof“Usefulness”forcommuters.
Step5:Weighteachbus line’sUsefulnessscoreaccording to theportionof the
studyarea’saffordablehousingstockthatthebuslineserves.
Ourinterestisincomparingcorridorsbytheaccessibilitytojobsprovidedto
thembypublic transit.Therefore, the “Usefulness” scoresofbus lines thatdonot
servetheentirecorridor(whichismostbuslines)mustbeweightedaccordingly.As
showninMap9,forinstance,busline#18(shownwitha0.5milebufferaroundit)
only serves a portion of the affordable housing stock in the Manor corridor. To
calculatethisportion,useSelectbyLocationtoselectthehousingthatintersectsthe
studyarea.UsetheStatisticsdialogueboxintheAttributeTabletosumthenumber
ofaffordableunitsinthestudyarea.Thiswillbethedenominatorintheproportion.
For the Manor corridor, this number is 1373.5. (Recall that the housing data
includesanestimatethat25%ofallunitsinsmallClassCbuildingsareaffordable.
This explains the possibility of fractions of units in the data.) Now use Selectby
Location to select the housing within 0.5 miles of the bus line. Use the Statistics
dialogue box to sum the quantity of selected housing units. This number is the
numerator. In the case of bus line #18 passing through the Manor corridor, this
numberis373.5.Dividethenumeratorbythedenominatorandentertheresultinto
18
Map9:The#18buslineservesonlyaportionoftheManorCorridor
IH 3
5
US
18
3
US 290
LAM
AR
CA
ME
RO
N
KOENIG
MARTIN LUTHER KINGFM 969
Esri, HERE, DeLorme, MapmyIndia, © OpenStreetMap contributors, andthe GIS user community
Created by: Cliff Kaplan | Fall 2014Data Sources: CapMatro; U.S. Census Bereau
NAD 1983 StatePlane Texas Central FIPS 4203 (US Feet)¯ 0 0.8 1.60.4 Miles
Legend
Units
Less than 50
51 - 100
101 - 200
More than 200
Jobs!( 20 - 300
!( 301 - 1000
!( 1001 - 5000
!( More than 5000
Bus line #18
0.5 mile buffer around bus line #18
Manor Corridor
19
Map 10: The #350 bus line serves only a portion of the Manor Corridor
IH 3
5
US
18
3
LAM
AR
US 290
CA
ME
RO
N
KOENIG
MARTIN LUTHER KINGFM 969
DE
SS
AU
Esri, HERE, DeLorme, MapmyIndia, © OpenStreetMap contributors, andthe GIS user community
Created by: Cliff Kaplan | Fall 2014Data Sources: CapMatro; U.S. Census Bereau
NAD 1983 StatePlane Texas Central FIPS 4203 (US Feet)¯ 0 0.8 1.60.4 Miles
Legend
Units
Less than 50
51 - 100
101 - 200
More than 200
Jobs!( 20 - 300
!( 301 - 1000
!( 1001 - 5000
!( More than 5000
Bus line #350
0.5 mile buffer around bus line #350
Manor Corridor
20
anewfieldintheattributetable;thisistheproportionbewhichwewillweighthe
“Usefulness” scores. Bus line #18 is useful to 373.5 / 1373.5 = 0.27, or
approximately 27% of the affordable housing in the corridor. Map 10 shows the
portionoftheManorcorridorthatisservedbybusline#350.
Figure 4: The same attribute table, now with the added Pn_of_Uts field(“Portion of Units Served”), which will serve as the weights for the“Usefulness”ofeachbusline.
RepeatSteps3,4,and5untiltheBusesatRushHour,Jobs,B_x_J(Usefulness),
and Pn_of_Unts (theweights) fields have been populated for all of the bus
lines.
Step6:Weighandsumthe“Usefulness”scores
AddanewfieldtotheAttributeTableandusetheFieldCalculatortomultiply
the“Usefulness”scorebytheassociatedweight(representingtheportionofunitsin
thecorridor that isable to takeadvantageof theusefulnessof thatparticularbus
line).
Figure5:Thesameattributetable,nowwiththeadded“WeightedUsefulness”field.
Sum these weighted scores. In the case of the Manor corridor, the sum is
587,336.93.
21
Step 7:Multiply the study area’s Usefulness score sum by the portion of all
transit‐accessiblejobsthataretransit‐accessiblefromthestudyarea.
WhenweaddedtogethertheweightedUsefulnessscoreforeachbusline,we
necessarilydoublecountedanyjobsthatareservedbymorethanonebusline(or
triplecountedthem,etc.).Wecanimagineacorridorthathastransitaccesstoalow
numberof jobs, but forwhich there aremanybuses that come frequently to take
commuterstothosejobs.Iftherearemanymoreworkerslivinginthecorridorthan
there are jobs that are accessible from the corridor, it is not as good as having
enough jobs for all theworkers.We account for this bymultiplying the summed
Usefulnessscoreforthestudyarea(587,336.93inthecaseofManor)bytheportion
of all transit‐accessible jobs thatare transit‐accessible from the studyarea.Aswe
calculatedearlier,thereare97,151jobsthataretransit‐accessiblefromtheManor
corridor.Thereare198,082jobsaccessiblebytransitcitywide,whichiscalculated
byusingSelectbyLocationtoselectallofthejobswithin0.5milesoftheentirebus
systemandusingStatisticsintheAttributeTabletosumthenumberofjobsinthe
relevantpayrange.Therefore,thefinalscorefortheManorcorridoris587,336.93x
[97,151/198,082]=288,064.39.
Thisscoreisacompositeofthenumberofaccessiblejobsfromeachbusline
multipliedby thenumberofbusesrunningon the line,weighted for theextent to
which the lineserves theaffordableunits in thecorridor,and the totalnumberof
jobs accessible by transit from the corridor. Therefore, the final score is best
understoodasanindexusefulforcomparingcorridorstooneanother,butwithout
aneasilygraspablemeaninginitsownright.
22
Results
I conducted the analysis described above on the Manor, Burnet, and
Riversidecorridors.Thefollowingtablesshowtheresultingnumbers.
Figure6:CalculationofresultsfortheManorcorridor
Figure7:CalculationofresultsfortheBurnetcorridor
23
Figure8:CalculationofresultsfortheRiversidecorridor
Figure9:Asummaryofthefinalscores
Corridor Final Score Actual number of accessible jobs
Manor 288,064.39
97,151
Burnet 1,195,604.92
198,082
Riverside 542,322.67
104,864
Analysis
According to this geospatial analytical method, affordable housing in the
Burnetcorridorhasslightlymorethantwotimesasmuchaccessibilityviatransitto
low‐payingjobsinAustinasdoesaffordablehousingintheRiversidecorridor,and
slightlymorethanfourtimesasmuchasthatintheManorcorridor.Thatresultis
mademoreinterestingbytheobservationthatthoserelationshipsarenotmirrored
24
bytheactualnumberofjobsaccessiblebytransitfromeachcorridor.Forinstance,
the Manor corridor’s final accessibility score is only 47% less than that of the
Riverside corridor.However, the actual numberof jobs accessibleby transit from
Manorisonly7%lessthanthenumberofjobsaccessiblebytransitfromRiverside.
Thedisparitybetweenthefinalaccessibilityscoresforthesetwocorridorsis
largelyexplainedbythefactthattheRiversidecorridorhasmuchhighervolumesof
bus traffic. 30 buses typically pass through theManor corridor during rush hour,
while54typicallypassthroughtheRiversidecorridor,accordingtoCapMetrobus
schedules.Thisdisparityisexacerbatedbythefactthat,onaverage,busespassing
throughtheManorcorridoronlyrunwithinaneasywalk(0.5miles)of46%ofthe
affordableunitsinthecorridor,whilethatfigureis57%intheRiversidecorridor.
But what accounts for the Burnet corridor’s score being twice that of the
Riverside corridor? Typically, 39 buses pass through the Burnet corridor during
rush hour, 20% fewer than pass through the Riverside corridor. Nor is coverage
remarkably better in the Burnet corridor than it is in the Riverside corridor; the
average bus passing through the Burnet corridor is accessible to 60% of the
affordableunitstherein,comparedto57%intheRiversidecorridor.
ThedifferenceinaccessibilitybetweentheBurnetcorridorandtheRiverside
corridorappearstobeexplainedbywherethebusesgo.Theaveragenumberofjobs
accessiblefromabusthatpassesthroughtheBurnetcorridoris50,145whereasthe
average number of jobs accessible from a bus that passes through the Riverside
corridoris31,678,37%fewer.The“Usefulness”scoresfortheaveragebuslinesthat
25
pass through the Burnet and Riverside corridors respectively are 265,860 and
184,266;Riverside’sscoreis31%lowerthanBurnet’s.
According to these observations, it appears that Manor’s low accessibility
score is caused by infrequent buses and a selection of buses that travels to a
relatively low number of jobs (97,151), while Burnet’s high accessibility score is
causedbythefactthateachbusservingtheBurnetcorridortendstotraveltomany
more jobs thando thebuses serving theother corridors, and together theBurnet
buses reach many more jobs (144,894) than do the buses serving the other
corridors.Maps11a,11b,and11cshowtheaffordablehousingstockineachstudy
corridor,asitrelatestobuslinesofvarying“Usefulness.”
Futureworkandconclusion
Continueddevelopmentofthisanalyticalmethodiswarranted.Conceptually,
manyofthemeasuresusedtocomparethecorridorstooneanother(“Usefulness,”
for example) are not as clear as they might be, in part due to their composite
natures.Itwillbeworthwhiletofindwaysofcombiningthedatathatcapturesthe
variation but is more intuitive to interpret. Technically, this method is time
consuming; its iterative nature (repeating several steps for each bus line in each
corridor) ought to make it automatable using Model Builder. I will likely pursue
bothconceptualandtechnicalstreamlines,asIcontinuetousethismethodforthe
Greenand InclusiveCorridorsproject.Thus far, the analysis shows that thereare
severalfactorsthatcancauseacorridor’shousingstocktohavemoreorlesstransit
access to job opportunities, and that these factors affect different corridors in
different ways. Planners seeking to preserve affordable housing with high
26
accessibilitytowork,orthoseseekingtoimproveaccessibilitywhereitis low,can
usethisanalyticalmethodtodiagnoseproblemsandstrategizeforsolutions.
Maps11a,11b,and11c:Theaffordablehousingstockineachstudycorridor,asitrelatestobuslinesofvarying“Usefulness.”Thethicknessofthebuslineindicatesthenumberofjobsthatareaccessiblefromthatline.Map11a:Manor Map11b:Burnet
Map11c:Riverside
27
ListofMaps
Map1:Thethreestudyareas–Page4
Map2:Thelocationsandsizesoftheaffordablehousingstock–Page7Map3:TheCapMetrobussysteminAustin,TX–Page8Map4:Jobswithinthe$15,000to$40,000annualpayrange–Page9Map5:ThebusesthatservetheManorcorridor–Page12Map6:JobsaccessiblebytransitfromtheManorcorridor–Page13Map7:Jobsaccessiblefromthebusline#18–Page15
Map8:Jobsaccessiblefromthebusline#350–Page16
Map9:The#18buslineservesaportionoftheManorCorridor–Page18Map10:The#350buslineservesaportionoftheManorCorridor–Page19Maps11a,11b,and11c:Theaffordablehousingstockineachstudycorridor,
asitrelatestobuslinesofvarying“Usefulness.”–Page26
28
DataAffordableHousingInventory[spreadsheet].(2014).BBCConsulting:Austin,TX.
AffordableHousingSurvey[spreadsheet].(2014).HousingWorks:Austin,TX.
CapitalMetro Bus Routes [shapefile]. (2012). CapMetro data and statistics portal.
Available:https://www.capmetro.org/datastats.aspx?id=129
Jobsthatpaybetween$15,000and$40,000annually[shapefile].(2011).U.S.Census
Bureau online interface, On the Map. Available:
http://onthemap.ces.census.gov/
Large Class C properties in Austin, TX [spreadsheet]. (2014). Investor Incentives:
Austin,TX.
ClassCmulti‐familyparcels[spreadsheet].(2014).TravisCountyAppraisalBoard:
Austin,TX.