geocoding and routing

23
CS 128/ES 228 - Lecture 9 a 1 Geocoding and Routing Ya can’t get thereah from hereah!

Upload: diata

Post on 07-Jan-2016

29 views

Category:

Documents


0 download

DESCRIPTION

Geocoding and Routing. Ya can’t get thereah from hereah!. The Traveling Salesman Problem (TSP). Insert Joke Here. A Salesman must visit each of N cities and return to his home base, driving as few miles as possible. Difficulty of TSP. Of interest to theoretical computer scientists because - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 1

Geocoding and Routing

Ya can’t get thereah from

hereah!

Page 2: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 2

The Traveling Salesman Problem (TSP)

A Salesman must visit each of N cities and return to his home base, driving as few miles as possible

Insert Joke Here

Page 3: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 3

Difficulty of TSP

Of interest to theoretical computer scientists

because

The Problem Is HARD

So hard that if you could compute an exact solution efficiently, you could also disrupt the world’s financial markets

Page 4: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 4

Applications of TSP

Inspection routes

Traveling Salespersons

Delivery professionals

Page 5: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 5

Knowing Where You Are

GEOCODING “is the process of converting an address into a point location.”

McDonnell and Kemp, International GIS Dictionary, 2nd Ed., 1998

“Hey, hey, hey. Don't be mean. We don't have to be mean because, remember, no matter where you go,

there you are.”

Buckaroo Banzai (from The Adventures of Buckaroo Banzai Across the Eighth Dimension)

Page 6: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 6

TIGER Maps – The Key to Routing

TIGER mapsare the primary data set used for geocoding.

… and geocoding is the basis for all routing.

Page 7: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 7

TIGER Map of the Area

Dr. Levine’s House

Page 8: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 8

TIGER Attribute Data

Page 9: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 9

Some Useful Fields

FID Field ID Number

LENGTH Length of full extent

FNODE Intersection of “start” of street

TNODE Intersection of “end” of street

FENAME Street name (root)

FRADDR # of first house on RIGHT side

TOADDL # of last house on LEFT side

ZIPR Zip code of RIGHT side

Page 10: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 10

How to Find an Address

1. Find the records for the appropriate street. (There may be more than one.)

2. Find the particular record whose address range includes your house number.

3. Decide “left” or “right” based on parity.

4. Interpolate.

23

Page 11: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 11

But Apply Georgian’s First Law

Don’t process the data yourself; get someone else to do it

(of data processing)

http://www.mapquest.com

Page 12: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 12

Limitations of Geocoding

Mapquest says Dr. Georgian lives here

Dr. Georgian says he lives here

Whom do you believe?

Page 13: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 13

Why Is It Wrong?

FRADDR 3179

TOADDR 3297

FRADDL 0

TOADDL 0

Where is Dr. Levine’s House?

(#3299)

Page 14: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 14

A Common TIGER Error

TIGER Files commonly miscode the F… and T… fields.

This results in various “mirroring” errors.

Newer versions of the data make this error less frequently.

Both Dr. Georgian and Dr. Levine are “victims” of this kind of error.

Page 15: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 15

Other TIGER Errors

This road does not exist!

Data Lag

causes errors as new roads are (or are not) built

Page 16: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 16

How Do We “Really” Do It?

By following instructions in tomorrow’s lab!

Page 17: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 17

Going “Big” Time

One can easily geocode many addresses at once.

Place them all in a table and then ask the system to do so.

This gives you all of your “hot” locations.

Page 18: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 18

Now, How Do You Visit Them All?

Page 19: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 19

How many ways are there?

For 20 stops, there are

20x19x18x…x3x2x1 =

2,432,902,008,176,640,000

Possible routes! (that’s 2 quintillion)

Once again, we cannot be exact

Once again, we need a GIS

Page 20: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 20

A Simple (non-)Solution Choose the route yourself

It turns out that humans, particularly experienced humans, are good at this

Then check only your 3 (or 40) best guesses!

Page 21: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 21

Working with Simpler Routes

So it can be done

And we can’t do it

Page 22: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 22

A Poor Man’s Solution

1. Select the lines that comprise your route.

2. Create a layer out of the selection.

3. Create a query that sums the lengths of the roads.

Page 23: Geocoding and Routing

CS 128/ES 228 - Lecture 9a 23

Conclusions

• Finding, Creating, and Measuring routes is a key task of many GIS’s.

• With ArcMap, we can do this for Points (geocoding), but not for extended routes.

• There are some “workarounds” we can use.

?