Introduction to Opportunity Mapping

1
Introduction to Opportunity Mapping

• OPPORTUNITY MAPPING WORKSHOP
• Nov. 30, 2007
• Samir Gambhir
• GIS/Demographic Specialist

2
Presentation overview

• SECTION I Introduction
• SECTION II Methodology
• SECTION III Data and analysis
• SECTION IV Future possibilities

3
Section Iintroduction
4
The community of opportunity approach

• Where you live is more important than what you
  live in
• Housing -- in particular its location -- is the
  primary mechanism for accessing opportunity in
  our society
• Housing location determines
• the quality of schools children attend,
• the quality of public services they receive,
• access to employment and transportation,
• exposure to health risks,
• access to health care, etc.
• For those living in high poverty neighborhoods,
  these factors can significantly inhibit life
  outcomes

5
Opportunity structures
6
framework

• The Communities of Opportunity framework is a
  model of fair housing and community development
• The model is based on the premises that
• Everyone should have fair access to the critical
  opportunity structures needed to succeed in life
• Affirmatively connecting people to opportunity
  creates positive, transformative change in
  communities

7
The web of opportunity

• Opportunities in our society are geographically
  distributed (and often clustered) throughout
  metropolitan areas
• This creates winner and loser communities or
  high and low opportunity communities
• Your location within this web of opportunity
  plays a decisive role in your life potential and
  outcomes
• Individual characteristics still matter
• but so does access to opportunity, such as good
  schools, health care, child care, and job networks

8
Opportunity mapping

• Opportunity mapping is a research tool used to
  understand the dynamics of opportunity within
  metropolitan areas
• The purpose of opportunity mapping is to
  illustrate where opportunity rich communities
  exist (and assess who has access to these
  communities)
• Also, to understand what needs to be remedied in
  opportunity poor communities

9
background (contd.)

• Evolved out of neighborhood indicators project
• Neighborhood Indicators
• Census 2000 data provided detailed neighborhood
  indicators
• Resulted in surge in neighborhood indicators
  based analysis
• Provided a snapshot of social and economic health
  of neighborhoods
• Shortcomings
• Each indicator is analyzed and mapped separately
• Overlay provides a complex view, hard to
  interpret

10
background (contd.)

• Opportunity mapping intended to provide a
  comprehensive view of any number of indicators

11
background (contd.)

• Resulted in a methodology that captures region
  wide opportunity distribution, in a comprehensive
  manner and it is reflective of todays
  metropolitan characteristics
• Ignores Urban-Suburban dichotomy
• Reflective of new trends decline of the inner
  suburbs, exurbs, inner city gentrification
• Reflective of the unique nature of each
  community e.g. Austin, TX vs. Cleveland, OH

12
background

• One of the major applications at Kirwan Institute
  was Chicago MSA opportunity classification (in
  collaboration with Institute on Race and Poverty,
  University of Minnesota

13
Section Iimethodology
14
Methodology

• Identifying and selecting indicators of
  opportunity
• Identifying sources of data
• Compiling list of indicators (data matrix)
• Calculating Z scores
• Averaging these scores

15
MethodologyIdentifying and Selecting Indicators
of High and Low Opportunity

• Established by input from Kirwan Institute and
  direction from the local steering committee
• Based on certain factors
• Specific issues or concerns of the region
• Research literature validating the connection
  between indicator and opportunity
• Central Requirement
• Is there a clear connection between indicator and
  opportunity? E.g. Proximity to parks and Health
  related opportunity

16
MethodologySources of Data

• Federal Organizations
• Census Bureau
• County Business Patterns (ZIP Code Data)
• Housing and Urban Development (HUD)
• Environmental Protection Agency (EPA)
• State and Local Governmental Organizations
• Regional planning agencies
• Education boards/school districts
• Transportation agencies
• County Auditors Office
• Other agencies (non-Profit and Private)
• Schoolmatters.org
• DataPlace.org
• ESRI Business Analyst
• Claritas

17
MethodologyIndicator Categories

• Education
• Student/Teacher ratio? Test scores? Student
  mobility?
• Economic/Employment Indicators
• Unemployment rate? Proximity to employment? Job
  creation?
• Neighborhood Quality
• Median home values? Crime rate? Housing vacancy
  rate?
• Mobility/Transportation Indicators
• Mean commute time? Access to public transit?
• Health Environmental Indicators
• Access to health care? Exposure to toxic waste?
  Proximity to parks or open space?

18
Methodologyeffect on opportunity

• Examples
• Poverty vs Income
• Vacancy rate vs Home ownership rate

19
MethodologyCalculating Z Scores

• Z Score a statistical measure that quantifies
  the distance (measured in standard deviations)
  between data points and the mean
• Z Score (Data point Mean)/ Standard
  Deviation
• Allows data for a geography (e.g. census tract)
  to be measured based on their relative distance
  from the average for the entire region
• Raw z score performance
• Mean value is always zero z score indicates
  distance from the mean
• Positive z score is always above the regions
  mean, Negative z score is always below the
  regions mean
• Indicators with negative effect on opportunity
  should have all the z scores adjusted to reflect
  this phenomena

20
MethodologyCalculating Opportunity using Z
Scores

• Final opportunity index for each census tract
  is the average of z scores (including adjusted
  scores for direction) for all indicators by
  category
• Census tracts can be ranked
• Opportunity level is determined by sorting a
  regions census tract z scores into ordered
  categories (very low, low, moderate, high, very
  high)
• Statistical measure
• Grounded in Social Science research
• Most intuitive but other measures can be used
• Example
• Top 20 can be categorized as very high, bottom
  20 - very low

21
Methodology Averaging Z scores

• Z score averages assume equal participation of
  all variables toward Opportunity Index
  calculations
• No basis to provide unequal weights
• Issue of weighting should be considered carefully
• Need to have a strong rationale for weighting
• Theoretical support would be helpful
• Arbitrary weighting could skew the results

22
Examples of opportunity mapping
23
Austin MSA, TX
24
New orleans msa, la
25
Baltimore msa,md
26
Ohioeducationopportunity
27
Cleveland msa,oh
28
Ongoing opportunity mapping projects

• Atlanta MSA, GA
• State of Massachusetts
• State of Connecticut

29
Section Iiidata and analysis
30
Data sources

• Census Data
• Non-Census Data

31
Census 2000 overview

• Information about 115.9 million housing units and
  281.4 million people across the United States
• Census 2000 geography, maps and data products are
  available
• Website www.census.gov

32
Geography hierarchy
33
Census 2000Short Form and Long Form
Short form
Long form
34
Short form

• 100-percent characteristics A limited number of
  questions were asked of every person and housing
  unit in the United States. Information is
  available on
• Name
• Hispanic or Latino origin
• Household relationship
• Race
• Gender
• Tenure (whether the home is owned or rented)
• Age

35
long form
?
For the U.S. as a whole, about one in six
households received the long-form questionnaire.
36
long form (contd.)

• Additional questions were asked of a sample of
  persons and housing units. Data are provided on
• Population

37
long form (contd.)

• Housing

38
Census 2010

• For Census 2010
• No long form questionnaire
• Short form questionnaire only
• To all residents in the U.S.
• Ask the same set of questions
• American Community Survey (ACS) to collect more
  detailed information
• Will provide data every year rather than every 10
  years
• Sent to a small percentage of population on a
  rotating basis
• No household will receive the survey more often
  than once every five years
• It might take at least five years, and some data
  aggregation, to get Census tract or smaller
  geography level data

39
Available short form data

• 100 data or short-form information
• Summary File 1
• Counts for detailed race, Hispanic or Latino
  groups, and American Indian/Alaska Native tribes
• Tables repeat for major race groups alone, two or
  more races, Hispanic or Latino, White not
  Hispanic or Latino
• Geography block, census tract
• Summary File 2
• 36 Population tables at census tract (PCT) level
• 11 Housing tables (HCT) at census tract (HCT)

40
Available long form data

• Sample data or long-form information
• Summary File 3
• 813 tables of data
• Counts and cross tabulations of sample items
  (income, occupation, education, rent and value,
  vehicles available)
• Lowest level of geography block group
• Summary File 4
• Tables repeated by race, Hispanic/ Latino, and
  American Indian and Alaska Native categories, and
  ancestry 336 categories in all.

41
Census basedmaps

• Fairly simple in calculations
• Easy to display
• Easy readability for the audience

42
Census data issues

• Historical data hard to get
• Inconsistent categories
• Block group and census tract boundaries are
  regularly updated
• Private data providers such as GeoLytics provide
  historical census data normalized to 2000
  geographies
• Inconsistency in data categories are minimized
  but still exist

43
Non-census data

• Data not available at census is gathered from
  other sources
• Good news!! It is available
• Bad news!! It might not be available at the
  geography of analysis (census tracts)
• Data needs to be manipulated to represent census
  tracts

44
Non-census dataExampleS

• School data
• Student poverty, test scores and teacher
  experience data might be available at
  school/District/County/State level
• Transit data
• Transit route data might be available with the
  local Metropolitan Planning Organization (MPO)
• Bus-stops or train stations might be available as
  a point theme
• Environmental data
• Toxic sites and toxic release data available at
  EPA as point data
• Parks and open spaces are available as shapefiles
• Public health
• Hospital locations might be available
• Main issue How to represent this data at census
  tract level

45
Spatial techniques

• Mapping software offers many techniques for data
  manipulation. Some of these methods used in our
  analysis are
• Interpolation
• Areal Interpolation
• Buffering

46
Interpolation

• Technique to predict value at unknown locations
  based on values at known locations
• Example Weather data
• Areal interpolation - Transferring data from one
  geography to another based on the proportion of
  area overlapping the target area
• Data aggregation
• Example - Transferring jobs data at zip code
  level to census tracts

47
buffering

• Buffering
• Creating a buffer of a specified radius around
  our data point
• Buffer distance decision should be research or
  knowledge based
• Captures proximity of events such as grocery
  stores, jobs etc.

48
Data issues and considerations

• Missing data
• Input data average
• Z score as zero
• Macro level data
• Jurisdictions or school districts
• When do we use ratio
• Grocery stores
• Jobs

49
Section Ivfuture possibilities
50
Future possibilities

• Web-based mapping
• Currently used mainly to display information
• Provides tools to zoom to scale, identify and
  some analysis
• Can be developed to exchange live information
• Google mash-up
• http//housingmaps.com
• http//wayfaring.com
• http//walkscore.com
• Mapping blogs
• Could residents go on-line and show where
  impediments to opportunity are in their
  neighborhood, or share their experiences?
• Semantic mapping
• Intelligence based Internet mapping