Taxis Are Our Friends - PowerPoint Presentation

Slide1

Taxis Are Our Friends

Mapping the “taxi-friendliness” of neighborhoods in the Westside of Los Angeles County

Earl Kaing

UP206A – Intro to GIS12/6/2011Final Presentation

Source: Earl

KaingSlide2

Introduction

Source: D.L.

ScrimgerSlide3

The Urban AgendaOnce someone is forced to buy a car, it’s all over: the private automobile is a huge investment; and once you sink money into that investment, the marginal costs (both real and perceived) of driving are almost negligible. In other words, when you own a car, there really is no incentive to seek out—or politically support—alternatives.

To move away from auto-dependency, we need to prevent that first purchase: If we want to move away from auto-dependency, we need to build political support for the kind of policies needed to make walking, bicycling, and public transit more viable alternatives. And to build this political support, we need to prevent that first purchase. We have to make it at least possible, if not easy, to live without owning a car.

The taxi industry makes it possible to live in an auto-centric world, without having to own your own car.

If we can expand the number and variety of trips that can be effectively served by taxis, the dramatic difference in quality of life separating the car-dependent from the car-free narrows. As the gulf narrows, more and more people are able to make that leap away from auto-dependency—to live rich and full lives on foot, by bike, on transit, and –every so often--in a taxi.The newly liberated expand the realm of what is politically possible: more compact, dense development; the widening of sidewalks; charging the right price for parking; policies which finally put people first; closing off downtown streets every single day of the week instead of once or twice a year!

The possibilities are endless. Slide4

Research GoalGoal : Expand the number and variety of trips that can effectively be served by taxis in Los Angeles, with the goal of supplementing—not replacing—trips on foot, bike, and transit

Taxicab Economics 101[Cost of Taxi Service] = f (distance, time, deadheading costs)In current system, customers pay a distance/time based rate that factors in an “average” deadheading cost—the cost of returning from a destination without passengers

Deadheading costs are a SIGNIFICANT! A 4-mile trip from Westwood to Bel-Air costs more for a taxi driver to serve than a 4-mile trip from Westwood to Santa Monica, but they are priced exactly the same! Midterm Research Question:

What if we could identify zones in Los Angeles where the deadheading costs are low? In other words, where the taxi driver is very likely to be able to find a return fare? Final Research Question: What would a network of taxi-friendly nodes in Los Angeles look like? Where would the nodes be located, and how much would it cost to travel between these nodes?

Imagine

getting picked up

in the center of Westwood Village and dropped

off in the middle of West

Hollywood—all for

$10

!!!Slide5

Neighborhoods of the Westside

18 neighborhoods

Generally bounded by the Pacific Ocean to the West, Fairfax to the East, the Santa Monica Mountains to the North, Manchester to the South

Average Median Household Income: $67,000

Intersected by two major highways

Map prepared by Earl Kaing

Data Source: LA Times, 2000 Census, LA County CIOSlide6

For the Midterm…

% Multi-Unit Housing Score

Median HH Income Score

Commercial Rent

Score

Commercial Density Score

Commercial Taxi Friendliness

Residential Taxi Friendliness

Maps prepared by Earl Kaing

Data Source: LA Times, 2000 Census, LA County CIO, LA County AssessorSlide7

For the Final…For the final, I will:Use Map Algebra

to consolidate the maps of residential and commercial taxi friendliness into a single mapUse Geocoding

to place a taxi stand at each area of taxi friendliness based on qualitative / experiential knowledge of areas that are pedestrian friendlyUse Service Area

Analysis Determine how many people live within 15 minutes walk, and 15 minutes bicycling of each taxi standCreate an O-D Matrix to estimate the expected fare between each taxi stand

1. Map Algebra

2.

Geocoding

3. Service Area

4. O-D MatrixSlide8

findings

Source: D.L.

ScrimgerSlide9

Taxi Friendliness ComponentsCommercial Taxi Friendliness

Residential Taxi Friendliness

[Residential Taxi Friendliness] = [% Multi-Unit Housing Quintile] + [Median HH Income Score*]

*See appendix for calculation

[Commercial Taxi Friendliness] = [Commercial Density Quintile] + [Commercial Rent Quintile]

1. Map Algebra

2.

Geocoding

3. Service Area

4. O-D Matrix

Maps prepared by Earl Kaing

Data Source: LA Times, 2000 Census, LA County CIO, LA County AssessorSlide10

Aggregate Taxi Friendliness

[Aggregate Taxi Friendliness] = [Commercial Taxi Friendliness] + [Residential Taxi Friendliness]

1. Map Algebra

2.

Geocoding

3. Service Area

4. O-D Matrix

Maps prepared by Earl Kaing

Data Source: LA Times, 2000 Census, LA County CIO, LA County AssessorSlide11

Taxi Stand Locations

1. Map Algebra

2.

Geocoding

3. Service Area

4. O-D Matrix

1

2

3

4

5

6

Map prepared by Earl Kaing

Data Source: LA Times, 2000 Census, LA County CIO, LA County Assessor

Image Sources : Google Street ViewSlide12

Service Area (Walking)

1. Map Algebra

2.

Geocoding

3. Service Area

4. O-D Matrix

Map prepared by Earl Kaing

Data Source: LA Times, 2000 Census, LA County CIOSlide13

Service Area (Bicycling)

1. Map Algebra

2.

Geocoding

3. Service Area

4. O-D Matrix

Map prepared by Earl Kaing

Data Source: LA Times, 2000 Census, LA County CIOSlide14

A Comparison of Estimated Fares Traditional Taxi

Awesome Taxi

1. Map Algebra

2.

Geocoding

3. Service Area

4. O-D Matrix

Map prepared by Earl Kaing

Data Source: LA Times, 2000 Census, LA County CIO

Fare Data: taxifarefinder.comSlide15

Potential Flat Fare Structure

1. Map Algebra

2.

Geocoding

3. Service Area

4. O-D Matrix

From

To

Flat Fare

SaMo

Venice

$5

SaMo

Westwood (Village)

$10

SaMo

Little Osaka

$10

SaMo

Westwood (Campus)

$10

SaMo

WeHo West

$15

SaMo

WeHo East

$20

From

To

Flat Fare

Venice

SaMo

$5

Venice

Little Osaka

$10

Venice

Westwood (Village)

$15

Venice

Westwood (Campus)

$15

Venice

WeHo West

$20

Venice

WeHo East

$25

From

To

Flat Fare

WeHo East

WeHo West

$5

WeHo East

Westwood (Village)

$15

WeHo East

Little Osaka

$15

WeHo East

Westwood (Campus)

$15

WeHo East

SaMo

$20

WeHo East

Venice

$25

From

To

Flat Fare

WeHo West

WeHo East

$5

WeHo West

Westwood (Village)

$10

WeHo West

Little Osaka

$10

WeHo West

Westwood (Campus)

$10

WeHo West

Venice

$20

WeHo West

SaMo

$20

From

To

Flat Fare

Westwood

Westwood (Village)

$5

Westwood

Little Osaka

$5

Westwood

WeHo West

$10

Westwood

SaMo

$10

Westwood

WeHo East

$15

Westwood

Venice

$15

From

To

Flat Fare

Little Osaka

WeHo East

$5

Little Osaka

Westwood (Village)

$5

Little Osaka

Westwood (Campus)

$10

Little Osaka

Venice

$10

Little Osaka

SaMo

$10

Little Osaka

WeHo

West

$15Slide16

Questions?

Source: D.L.

ScrimgerSlide17

appendixRequirements ChecklistSlide18

Requirements Checklist

Requirement

How Met?

8 Layouts: Does presentation include a minimum of 8 layouts?Communities of the Westside (Slide 4)

For the Midterm (Slide 5)

Taxi Friendliness Components (Slide 7)

Aggregate Taxi Friendliness (Slide 8)

Taxi Stand Locations (Slide 9)

Service Area Walking (Slide 10)

Service Area Biking (Slide 11)

A Comparison of Estimated Fares (Slide 12)

7 Layers:

Does at least one layout include seven (7) or more layers?

Service Area Biking (Slide 11)

California Shoreline

Communities of the Westside

Major Highways

Tiger Roads

Taxi Stand Locations

Service Area Layer (5 min)

Service Area Layer (10 min)

Service Area Layer (15 min)

Modeling:

Does your presentation use a model to automate data manipulation? Is this model diagram included as a jpg at the end of the presentation or following the layout it was used in?

To create “Aggregate Taxi Friendliness” (Slide 8), I used a model to 1) convert the four components of taxi friendliness (2 residential and 2 commercial) into separate

rasters

, and then 2) reclassify each of these

rasters

into an indexed score from 1-5. A screenshot of this model can be found in the appendix.Slide19

Requirements Checklist

Requirement

How Met?

Metadata: Does your project include at least one metadata sheet for at least one of your original geographic layers or elements? Is the screenshot of this metadata sheet included at the end of the presentation?I created a metadata sheet for the Communities of the Westside shape file that I created. The screenshot of the metadata sheet can be found in the appendix.

Measurement/Analysis:

Does your project include a measurement analysis that integrates some measure of distance (buffer, concentric zones, elements displayed a certain distance from a central feature, nearest neighbor, or display lines/circles a given distance from a feature, etc)?

I used

ArcGIS

Network Analyst to

calculate walking and biking service areas for each taxi stand node based on 1) tiger roads distance; 2) average walking speed; and 3) average biking speed. The service areas illustrate temporal distance from each taxi stand.

Original Data:

Does your project include an original map layer created using data from outside sources?

I used

georeferencing

and feature editing to create the “Communities of the Westside” layer seen in Slide 4. In the midterm, I used a different

shapefile downloaded from the LA County GIS portal. For the final, I took the

shapefile I used in the midterm, and edited the features to match a georeferenced screenshot (JPEG) of Mapping LA’s “Communities of the Westside” page.

Descriptive Map: Does your powerpoint include a descriptive map that provides a general overview of your study area?The “Communities of the Westside” (Slide 4) provides a general overview of the study area.Slide20

Requirements Checklist

Requirement

How Met?

Six Additional Skills: Does your project utilize at least six other skills, one of which is drawn from the following? Extracting information from a buffer

Charts, graphs, or images

Hotspot analysis

Network analysis

Spatial analysis

Elevation

3-d modeling

Google Mash-Up

Charts, Graphs & Images:

To help give the audience a better feel for the built environment around each taxi stand location I integrated images from Google street view for each location into the layout of “Taxi Stand Locations” (Slide 9).

Network Analysis (Service Area):

I used network analyst to calculate the 5, 10, and 15 minute service areas around each taxi stand, for both the walking and biking modalities.

Network Analysis 2 (O-D Matrix):

I used network analyst to generate a matrix of network travel costs (in minutes) from each taxi stand location to all other taxi stands. I then used this matrix to estimate the dollar cost of service, based on the assumption that under this new system, service costs can be cut in half.

Hotspot Analysis:

I used hotspot analysis to create the “Aggregate Taxi Friendliness” layout (Slide 8) by calculating the intersection of the two commercial and two residential taxi friendliness factors. Slide21

Requirements Checklist

Requirement

How Met?

Six Additional Skills (cont.): Does your project utilize at least six other skills, one of which is drawn from the following? Extracting information from a buffer

Charts, graphs, or images

Hotspot analysis

Network analysis

Spatial analysis

Elevation

3-d modeling

Google Mash-Up

Extracting Information From a Buffer:

To calculate the total population within 15 minutes biking, and within 15 minutes walking of each taxi stand, I:

dissolved the 5, 10, and 15 minute service areas for each modality into a single feature (the buffer)

performed a spatial join between the buffer and the underlying census tracts (to which population counts had been joined) to sum up the population of all census tracts intersecting the buffer

estimated the population within the buffer as the proportion of the area of the buffer to the total area of all intersecting census tracts

Inset Map:

Used in Slide 4 (Communities of the Westside) to show the Westside Region in the context of Los Angeles County. Also used in Slide 8 (Aggregate Taxi Friendliness) to help transition from a higher level of zoom to a lower level of zoom.

Line Graduated Symbol:

Used in Slide 12 (A Comparison of Estimated Fares) to distinguish between low cost trips ($0-$15), in green; medium cost trips ($15-30), in yellow; and high cost trips ($30-$50), in red.Slide22

Requirements Checklist

Requirement

How Met?

Six Additional Skills (cont.): Does your project utilize at least six other skills, one of which is drawn from the following? Extracting information from a buffer

Charts, graphs, or images

Hotspot analysis

Network analysis

Spatial analysis

Elevation

3-d modeling

Google Mash-Up

Creating Indices:

created an aggregate taxi friendliness indicator by combining the residential and commercial taxi friendliness scores from the midterm, without any weights. The residential taxi friendliness = f(% multi-unit housing, median HH income). The commercial taxi friendliness = f(commercial parcel density, commercial rent ($/

sqft

) ).

Geocoding

:

to identify the taxi stand locations seen in Slide 9, I started at areas with high taxi friendliness scores, and then used my experiential knowledge and Google Street View to identify specific cross streets which would be ideal for a taxi stand. I then

geocoded

these intersections, using an address locator that I created based on the tiger roads shape file.Slide23

appendixStep-by-Step MethodsSlide24

Communities of the Westside

Took

a JPEG of Westside Region from Mapping LA website

Georeferenced

JPEG to give it coordinates

Used

georeferenced

as basis to create a new

shapefile

by editing the unofficial LA County communities

shapefile

to match the Mapping LA JPEG

Included the neighborhood of West Hollywood in my definition of the Westside, even though it’s not included by the Mapping LA project

Used

new shape file to determine which census tracts to consider in analysis. Any census tracts which intersected a Westside neighborhood was

included. All other tracts were clipped away.Slide25

Map Algebra

Combine the separate residential and commercial taxi friendliness

maps into a single taxi friendliness map.

Used

a model to convert residential shape file and commercial shape file into four separate

rasters

Used model to

reclass

each

rasters

. All were

reclassed

based on quintiles, with the exception of income, which I

reclassed

based on standard deviations from the average median income on The Westside

Used Map Algebra > Raster Calculator to add the two residential

rasters

to get a residential index. Repeated process with the commercial

rasters

to get a commercial index.

Added the two

rasters

together to create an aggregate taxi friendliness index

1. Map Algebra

2.

Geocoding

3. Service Area

4. O-D MatrixSlide26

Model

1. Map Algebra

2.

Geocoding

3. Service Area

4. O-D MatrixSlide27

Metadata

1. Map Algebra

2.

Geocoding

3. Service Area

4. O-D MatrixSlide28

Taxi Stand Locations

For each neighborhood, determine the best location to place a taxi stand.

Using the raster of taxi friendly census tract, I classified out of a total possible friendliness score of 20, those in the 90-100%, 80-90%, 70-80%, and 60-80% range.

Based on the raster, I identified unique clusters of 90-100% taxi friendliness within each neighborhood. Most neighborhoods had one distinct cluster, but some, like Santa Monica, had two.

I used Google maps, along with qualitative and experiential knowledge to identify specific cross streets for the taxi stands. I was looking for locations that were human-scale and pedestrian friendly.

Based on this analysis, I identified the following areas:

Santa Monica: SMB & 4

th

Sawtelle

:

Sawtelle

and Olympic

Venice: Abbot Kinney & Westminster

West Hollywood: San Vicente & SMB; Martel & SMB

Westwood:

Weyburn

and Broxton; Westwood and Strathmore

I used the

Tigerroads

shape file for Los Angeles, clipped to the Westside, and created an address locator based on it. The roads have dual ranges.

I then used this address locator to

geocode

the locations I had identified as most appropriate for a taxi stand.

For this layout, I included pictures of the intersection where the stand will be located, for visual reference.

1. Map Algebra

2.

Geocoding

3. Service Area

4. O-D MatrixSlide29

Service AreaCreate a service area to see who is 5, 10, 15 minutes away from the stand on foot, and by bike. I calculated a segment length for each road feature in the tiger roads shape file.

I then calculated impedence = [length] / [speed] for driving, biking, and walking, where I assumed:

average driving speed across the entire network of 25 mph (DMV speed limit in all business/residential districts unless otherwise posted)average walking speed of 3 mphaverage biking speed of 15 mph

Next, I created a network dataset using the updated tiger roads fileI then used spatial analyst to create a service area analysis layer for walking and biking. What area is within 5, 10, and 15 min walking or biking of the taxi stand?How many people live within 15 minutes walk or bike of each taxi stand? Dissolve the 5, 10, and 15 minute service areas into a single buffer layer.

Join the buffer layer with the census data layer containing information about population per census tract

Extract information based on spatial location, to sum up the population of all census tracts which intersect the buffer

Estimate the population within the buffer only using a factor = [area of buffer] / [total area of all census tracts which intersect buffer]

Repeat this for both the walking and the biking service area.

Represent taxi stand access with graduated symbols based on population served.

1. Map Algebra

2.

Geocoding

3. Service Area

4. O-D MatrixSlide30

O-D MatrixEstimate cost of service between each nodeUse network analyst to calculate an O-D matrix for the network.

If we assume current prices are twice as high as they need to be because of deadheading, then the new rate per unit time/distance for this new proposed schematic can be divided by two I use taxi fare calculator available online to see what the rate would be under current price regime. It turns out taxi trips average about $1.6 per minute. Thus the new price would be $0.8 per minute.

Calculate the new cost, using the driving time (minutes) between each node from the O-D matrix

1. Map Algebra

2.

Geocoding

3. Service Area

4. O-D Matrix