METRO Park-and-Ride Survey Presentation (Daniel Christen Graduate Thesis/Applied Project)

METRO Park-and-Ride Survey

Daniel J. ChristenApril 2011

SURVEY BACKGROUND

Survey Introduction• Purpose of the survey was to understand the travel patterns and characteristics

of Valley Metro Rail (METRO) park-and-ride users.

• Survey included 18 questions• Origin-destination locations• Opinions about METRO park-and-ride facilities• Personal socio-economic characteristics

• Eight of the 18 questions were looked at in detailed analyses• Respondent zip code locations (spatial distribution)• Respondent park-and-ride facility used (park-and-ride comparison)

• A commute time analysis was conducted to determine the time difference between direct commute (origin to destination) and park-and-ride commute (origin to destination via park-and-ride) for all respondents

Survey Distribution• Flyers were used to invite participants to fill out the survey on Survey Monkey.

The distributed flyers targeted all park-and-ride users among the eight METRO park-and-ride facilities within a given time frame.

• 2,500 flyers were used for distribution• 1,844 flyers were distributed on Tuesday, November 16, 2010, between 9:00

AM and 11:00 AM (proportionally)• 656 flyers not used on November 16 were distributed on Wednesday,

November 17, 2010, between 9:00 AM and 12:00 PM (equally)

• Flyers were distributed to 52.5% of all parking spaces on November 16, which represents total park-and-ride utilization

• 184 respondents based on the 2,500 flyers, thus a response rate of 7.36%

• Only 156 respondents defined both their origin and destination locations, thus the remaining 28 respondents were not included in the survey analyses.

Park-and-Ride Respondents

Flyer and Respondent Percentages

RESPONDENT DISTRIBUTION

QUESTION RESULTS

Detailed Questions• Eight of the 18 questions were looked at in detailed analyses

• Respondent zip code locations (spatial distribution)• Respondent park-and-ride facility used (park-and-ride comparison)

• The questions include…• Question 2: Weekly park-and-ride usage• Question 5: Perceived travel time from origin to park-and-ride• Question 7: Trip purpose (school vs. work)• Question 9: Parking availability• Question 10: Park-and-ride and light rail convenience• Question 11: Implementation of a park-and-ride fee• Question 16: Park-and-ride user age• Question 18: Park-and-ride user annual gross income

Types of Analysis Weights• Two types of analysis weights were used to solve for zip code or park-and-ride

facility values based on respondent attributes per question.• Linear weights• Weighted percentages

• Questions 2, 5, 16, and 18 were solved with linear weights because the questions could be scaled linearly (Ex. 1, 2, 3, 4, 5, 10, 20, …)• Weights were multiplied by the number of respondents per question

category per zip code or park-and-ride and then added to get a total value

• Questions 7, 9, 10, and 11 were solved with weighted percentage values because the questions had only two extreme values (Ex. No = -1, Yes = 1)• Weights were multiplied by the number of respondents per category, divided

by the total number of respondents (per unit of analysis), and then added to get a deviation value from “0”

More About Weighted Percentages

• Questions with weighted percentages that had more than two categories were given half weights of -0.5 and 0.5 depending on the relationship of the categories. (Ex. Never available = -1, Somewhat available = -0.5, Available = 0.5, and Always available = 1)

• Categories given a weight of -0.5 or 0.5 will only have half the influence of those categories with weights of -1 or 1.

• EXAMPLE: Four respondents in a zip code. Three respondents answer always available, while the fourth respondent answers available.• (3 respondents) x (weight of 1) = 3 / (4 total respondents) = 0.75• (1 respondent) x (weight of 0.5) = 0.5 / (4 total respondents) = 0.125• 0.75 + 0.125 = 0.875 deviation value

Question 2: Weekly Usage

Question 2 Park-and-Ride Results

Question 5: Perceived Travel Time

Question 5 Park-and-Ride Results

Question 7: Trip Purpose

Question 7 Park-and-Ride Results

Question 9: Parking Availability

Question 9 Park-and-Ride Results

Question 10: Convenience

Question 10 Park-and-Ride Results

Question 11: Parking Fee

Question 11 Park-and-Ride Results

Question 16: User Age

Question 16 Park-and-Ride Results

Question 18: User Income

Question 18 Park-and-Ride Results

COMMUTE TIME ANALYSIS

Analysis Background

• An analysis based on route distance and route time data was conducted in order to find the difference in commute times between origin and destination (direct commute), and between origin and destination via park-and-ride (park-and-ride commute).

• Three major components…• Origin Route Time: Time between origin location and park-and-ride• Destination Route Time: Time from park-and-ride to closest destination

station• Direct Route Time: Time between origin location and destination location

• The direct route time is the direct commute time, while the sum of the origin route time and destination route time is the park-and-ride commute time

Solving for Origin Route Time• Origin route time was solved by finding a common factor for all respondent

origin routes based on average distance and average travel time.• Average distance of all respondent origin routes was 6.43 miles• Average travel time was based on Question 5 category parameters

• Time factor estimates were compared to each other based on the amount of travel time values within corresponding stated travel time parameters for each respondent in Question 5. The time factor with the most respondents within stated parameters was 16.5 minutes for every 6.43 miles traveled.

• EXAMPLE: If a respondent answered 10 to 15 minutes for Question 5 and had a travel time of 12.25 minutes, where for every 17 minutes, 6.43 miles would be covered, that respondent would be “in parameters” for a time factor of 17 minutes.

• ORIGIN ROUTE TRAVEL TIME:6.43 miles / 16.5 minutes = (respondent origin route distance in miles) / X min.

Time Factor Comparison

Solving for Destination Route Time

• Destination route time for each respondent was solved by finding the light rail travel time between a respondent’s park-and-ride facility and the closest light rail station to their destination.

• Light rail travel times were solved using the METRO light rail schedule that become effective on January 24, 2011.

• The light rail station times for the train starting closest to 9:00 AM were used as the benchmark light rail times• Monday Through Thursday Eastbound: Starts at 8:57 AM at Montebello and

ends 10:02 AM at Sycamore• Monday Through Thursday Westbound: Starts at 9:06 AM at Sycamore and

ends 10:11 AM at Montebello

• The difference between the two station times per respondent is the destination route time

Solving for Direct Route Time• Solved using trigonometry (Law of Cosines):

• All three route distances are known (origin, destination, and direct)• Two of the three route times are known (origin and destination)

• NOTE: All three routes must be straight lines in order for trigonometry to work!• Why origin route distance was not based on the shortest path route• Straight destination route distance (park-and-ride to destination location) is

used instead of the real destination route distance of the light rail line

• Two steps of trigonometry are needed to solve for direct route time…• STEP 1: Solving for the angle of distance between the origin route and

destination route based on the known route distances• STEP 2: Angle measurement in radians is used to find the direct route time

based on the other two known route times

Destination Route Regression• A regression model showed straight destination route distances were statistically

similar to real light rail destination route distances at the 95% confidence level with an r-square value of 0.97, thus significant to use.

Trigonometry Steps 1 and 2

STEP 1 STEP 2

Park-and-Ride Average Time Values

Commute Difference

CONCLUSION

Significant Findings• FIRST: The three closest park-and-rides to ASU (Dorsey, McClintock, and Price)

had a significant amount of school purpose respondents compared to the other park-and-rides

• SECOND: Dorsey was the only park-and-ride in which respondents felt not enough parking existed at the facility

• THIRD: West Valley park-and-rides were considered slightly more convenient compared to East Valley park-and-rides

• FOURTH: Contrasting relationship between terminal park-and-rides (Montebello and Sycamore) and central park-and-rides (38th Street and Dorsey)• WEEKLY USAGE• USER AGE• TRAVEL TIME• COMMUTE DIFFERENCE

A Successful Survey Study?

• FIRST: Ratio between distributed flyers per park-and-ride and number of respondents per park-and-ride was similar, thus creating an accurate response rate per park-and-ride facility

• SECOND: Spatial distribution of respondents and respondent area shapes were similar to those found in other park-and-ride studies• 50% of respondents are within 5 miles of park-and-ride: SURVEY HAD 58.3%• 80% of respondents are within 10 miles of park-and-ride: SURVEY HAD 82.7%• Respondent areas (catchment areas) tend to be either parabolas or ellipses:

SURVEY HAD BOTH PARABOLAS AND ELLIPSES

• THIRD: Dorsey had the longest perceived travel time of 14.44 minutes, while having the largest commute difference in which the park-and-ride commute was 25.83% longer than the direct commute. Thus the commute time analysis a success!

THANK YOU