Introduction to Geographic Information Systems Spring 2013 (INF 385T-28437) Dr. David Arctur Lecturer, Research Fellow University of Texas at Austin

Introduction to Geographic Information Systems Spring 2013 (INF 385T-28437)

Dr. David ArcturLecturer, Research Fellow

University of Texas at Austin

Lectures 8 & 9Feb 28, 2013

8 - Spatial Analysis9 - Geocoding

ArcInfo coverages (from Lecture 5) Created using ESRI’s ArcInfo software

(prior to version 8) Older format (import/export as “.e00”) Set of files within a folder or directory called

a workspace Files represent different types of topology or

feature types Coverages have geometry: Arcs (lines), Nodes

(points), or Polygons, and associated attribute tables

Coverages also have Tics (spatial registration points), and may have Labels and Annotation

INF385T(28437) – Spring 2013 – Lecture 5

Review

Inside a coverage…


View from the operating system:

Coverage attribute table Area and perimeter

Coverage_ and Coverage_ID

4INF385T(28437) – Spring 2013 – Lecture 5

5

Labels vs. Annotation Labels are based on one or more attributes of features. Annotation is a way to store text to place on your maps

independent of features. Each piece of text stores its own position, text string, and display properties. Annotation can also be linked to individual features, for positional or existence dependency.

If the exact position of each piece of text is important, you should store your text as annotation in a geodatabase. Annotation provides flexibility in the appearance and placement of your text because you can select individual pieces of text and edit them.

You can convert labels to create new annotation features.INF385T(28437) – Spring 2013 – Lecture 8

Spatial Analysis Outline (Tutorial Ch.9)

Proximity buffers Site suitability example Basic apportionment (on your own) Advanced apportionment (on your

own)

Then… Geocoding (Tutorial Ch.7)


Lecture 8

PROXIMITY BUFFERSLecture 8


Proximity buffers Points

Circular buffers with user supplied radius

Lines Looks like worm based on line feature


Proximity buffers Polygons

Extends polygons outward and rounds off corners

Created by assigning a buffer distance around polygon


Point buffer example Polluting company buffers

Added schools Added population

INF385T(28437) – Spring 2013 – Lecture 8 10

Point buffer example Crimes near schools


Line buffer example Businesses within .25 miles of a

selected street


Select features in buffer


Spatial join to count Join business points to buffer polygon


Polygon buffer example River buffer to analyze environmental

conditions, flooding, etc.


Polygon buffer example Parcels within 150′ of selected property


Select features in buffer


SITE SUITABILITYLecture 8


Locate new police station Criteria

Must be centrally located in each car beat (within a 0.33-mile radius buffer of car beat centroids)

Must be in retail/commercial areas (within 0.10 mile of at least one retail business)

Must be within 0.05 mile of major streets


Starting map Lake Precinct of the Rochester, New

York, Police Department Police car beats Retail business points Street centerlines


Create car beat centroids XY centroids for police beats


Buffer car beat centroids .33 mile buffer


Buffer retail businesses 0.1 mile buffer


Select major streets Select by attribute


Buffer major streets 0.05 mile buffer


Intersect buffers Can only intersect two at a time

Car beat and businesses Streets


Site suitability result Map showing possible sites for police

station


Spatial Analysis Summary Proximity buffers (Tutorial exercise 9-1) Site suitability example (Tutorial exercise

9-2) Basic apportionment (optional) Advanced apportionment (optional)

Assignments: 9-1, 9-2 (9-3 optional)Next up today - Geocoding



BASIC APPORTIONMENTLecture 8


Apportionment example Population by voting district

You want to know the population of a voting district but only have census tracts

Voting districts and census tracts are not contiguous

Approximate the population of voting using census tracts and blocks


Population by voting district Start with census tracts


Population by voting district Overlay voting districts (not contiguous

with tracts)


Population by voting district Better to use block centroids for population

Smaller than tracts


Spatially join centriods Join centroids to voting districts


Other simple apportionments Population by

Neighborhoods Zip Codes Historic sites Others?


Census data to apportion Short form SF1 data (tract, block group,

block) Population Age Race Housing Units Others?

Long form SF3 data (tract and block group) Educational attainment Income Poverty status Others? 37INF385T(28437) – Spring 2013 – Lecture 8

ADVANCED APPORTIONMENTLecture 8


Advanced Apportionment Chapter 9 example

Police want to know the number of under-educated persons in their car beats

Under-educated data is located SF3 tables, census tracts or block groups (not car beat polygons)


Data to apportion Car beats Census tracts Beats and tracts

Not contiguous


Beats and tracts zoomed Tracts clearly cut across beats


Tract attribute table Tracts contain undereducated data

No high school degree


Math of apportionment Simple census data (e.g. population) is

not a problem Can use block centroids

Problem Block centroids don’t

contain undereducatedpopulation

Tracts contain thisinformation


Math of apportionment Tract 360550002100 Car beats 261 and 251


Math of apportionment One approach

Assume that the target population is uniformly distributed across the tract

You could split undereducated population up by the fraction of the area of the tract in each car beat

What if, however, the tract has a cemetery, park, or other unoccupied areas? Then the apportionment could have sizable errors


Math of apportionment A better approach

Use a block-level, short-form census attribute as the basis of apportionment

Assume that the long-form attribute of interest is uniformly distributed across the short-form population (accounts for unoccupied areas)

One limitation of the block-level data is that the break points for age categories do not match those of the educational attainment data (persons 25 or older)

The best that can be done with the block data is to tabulate persons aged 22 or older

Close enough for approximationINF385T(28437) – Spring 2013 – Lecture 8 46

Math of apportionment Tract 360550002100 has 39 block

centroids that span 2 beats

Of the 26 blocks making up the tract, the 13 that lie in car beat 261 have 1,177 people aged 22 or older.

The other 13 blocks in car beat 251 have 1,089 such people for a total of 2,266 for the tract.


Math of apportionment Apportionment assumes that the

fraction of undereducated people aged 25 or older is the same as that for the general population aged 22 or older This fraction, called the weight, is 1,177 ÷

2,266 = 0.519. For the other car beat, the weight is 1,089 ÷ 2,266 = 0.481

Thus, we estimate the contribution of tract 36055002100 to car beat 261’s undereducated population to be (1,177 ÷ 2,266) × 205 = 106. For car beat 251, it is (1,089 ÷ 2,266) × 205 = 99


Math of apportionment Eventually, by apportioning all tracts,

we can sum up the total undereducated population for car beats 261 and 251


BACKGROUND STEPSLecture 8


Background steps1.) Download census data

Download census block and tract polygons from the census Web sites for the county containing the administrative area polygons

Download the short-form census data for blocks that are the basis of apportionment, in this case the population of age 22 and greater

Download the long-form census attribute(s) at the tract level that you wish to apportion to the administrative area, in this case the population aged 25 or greater with less than high school education


Background steps2.) Create new tract layer

That intersects administrative boundaries

If a tract is only partially inside the administrative area, you must include the entire tract for apportionment to work correctly

An example tract is the southerly-most tract in Tutorial9-3.mxd


Background steps3.) Prepare block centroids

Create a new centroid point layer for blocks Clip the centroids with the new intersected tract

layer Join census short-form data to the clipped block

centroids This is the layer that is the basis for apportionment


Background steps4.) Sum the short-form census attributes in age

categories to create Age22Plus in the clipped block centroids table This step is unique to this problem Also, this table has a new TractID attribute which

concatenates FIPSSTCO & TRACT2000 to create an ID matching the Tracts map layer


Background steps5.) In the attribute table for block centroids,

sum the field for persons aged 22 or older by TractID to create a new table, SumAge22Plus. This table provides the denominator for the weight used in apportionment


APPORTIONMENT STEPSLecture 8


Apportionment steps1.)Intersect tracts and car beats to create new

polygons that each have a tract ID and car beat number


Apportionment steps2.) Spatially join the new layer of tracts and car

beats with the block centroids to assign all the tract attributes (including the attribute of interest: undereducated population) and car beat attributes to each block’s centroid


Apportionment steps2.)


Apportionment steps3.) Join SumAge22Plus to block centroids to

make the apportionment weight denominator, total population aged 22 or older by tract, available to each block centroid


Apportionment steps3.) Export the join as a precaution


Apportionment steps4.) For each block centroid, create new fields to

store apportionment weight and apportioned undereducated population values, then calculate these values


Apportionment steps4.) Calculate values


Apportionment steps5.) Sum the apportionment weights by tract as

a check for accuracy (they should sum to 1.0 for each tract)


Apportionment steps5.) Each tract that is totally within car beats will

have weights summing to 1. Those partially within car beats sum to less than 1


Apportionment steps5.) Sum the undereducated population per car

beat


Join apportionment results The last task is to join

the table containing undereducated population by car beat to the car beats layer, then symbolize the data for map display


Finished map


Summary Proximity buffers Site suitability example Basic apportionment Advanced apportionment


Introduction to Geographic Information Systems Spring 2013 (INF 385T-28437) Dr. David Arctur Lecturer, Research Fellow University of Texas at Austin

Documents

bufferinf385t28437 spring

schoolsinf385t28437

15inf385t28437 spring

76inf385t28437 spring

coverageinf385t28437

annotation inf385t28437

polygon 9inf385t28437

id 4inf385t28437 spring