FSUTMSVOYAGER MODELING TRAINING SERIES FSUTMS Introduction to Modeling

FSUTMS/VOYAGER MODELING TRAINING SERIES FSUTMS Introduction to Modeling Workshop Forecasting and Trends Office Intermodal Systems Development Florida Department of Transportation

Agenda Ø Workshop Introduction Ø What is a “Travel Demand Model”? Ø Who uses travel demand models and what are they used for? Ø How do I know if I have a good model? Ø What do I need? 2

Workshop Objectives Ø Provide the participant grounding in the fundamentals of transportation modeling under FSUTMS*. Ø Provide hands-on experience analyzing real-world problems using the FSUTMS tools, utilities, and processes in transportation modeling. *FSUTMS, Florida Standard Urban Transportation Model Structure, is a computerized transportation planning model package developed by the Florida Department of Transportation and Florida’s Model Task Force. 3

Instructor Roberto Miquel Senior Planner miquelro@cdmsmith. com 4

FDOT Project Management Terry Corkery FDOT Forecasting and Trends Office 605 Suwannee Street, MS 27 Tallahassee, FL 32399 -0450 (850) 414 -4903 Terrence. Corkery@dot. state. fl. us 5 Thomas Hill Systems Transportation Models Manager Florida Department of Transportation 850 -414 -4924 Thomas. Hill@dot. state. fl. us

FDOT District Modeling Coordinators Ø District 1 – Christopher Simpron (863) 519 -2343 christopher. simpron@dot. state. fl. us Ø District 2 – Ameera Sayeed (904) 360 -5647 ameera. sayeed@dot. state. fl. us Ø District 3 – Linda Little (850) 330 -1217 linda. little@dot. state. fl. us Ø District 4 – Hui Zhao (954) 777 -4635 hui. zhao@dot. state. fl. us Ø District 5 – Jason Learned (386) 943 -5320 jason. learned@dot. state. fl. us Ø District 6 – Neil Lyn (305) 470 -5373 neil. lyn@dot. state. fl. us Ø District 7 – Kenneth Spitz (813) 975 -6550 kenneth. spitz@dot. state. fl. us Ø District 7 – Andrew Tyrell (813) 975 -6458 andrew. tyrell@dot. state. fl. us Ø Turnpike – Cesar Segovia (407) 264 -3153 cesar. segovia@dot. state. fl. us 6 12

On-Line Workshop Opportunities Ø http: //www. fsutmsonline. net/online_training/index. html Ø FSUTMS Executive Summary Modeling Seminar Ø FSUTMS Comprehensive Modeling Workshop Ø FSUTMS Introduction to Cube Cloud Ø FSUTMS Webinar Series § § § 7 Model Reporting: Understanding the Performance Measures Reported from a TDM Origin and Destination Skims and Matrix Information Destination Choice – A Modeling Primer Friction Factors, BPR Curves, Speed, and Trip Length The Latest Land Use Models Available in Florida and Best Practices Data Resources for Validating and Calibrating TDMs in Florida

Other Potential Live Training Webinars Ø Introduction to R Programming 8

FY ’ 17 In-Person Training Schedule Ø May 23, 2017: Introduction to Modeling Workshop, Miami Ø May 24, 2017: Introduction to R Programming Workshop, Miami Ø June 21, 2017: Introduction to Modeling Workshop, Orlando Ø June 22, 2017: Introduction to R Programming Workshop, Orlando Ø June 28, 2017: Introduction to Modeling Workshop, Tallahassee Ø June 29, 2017: Introduction to R Programming Workshop, Tallahassee 9

Continuing Education Credits Ø P. E. § Include P. E. number on Attendee list and FDOT will process for you 10

Knowledge Sharing Opportunities Ø Florida Full Model Task Force (MTF) (annually) Ø MTF Committees § Model Advancement § Data / Geographic Information Systems (GIS) § Transit & Rail § Freight 11

Knowledge Sharing Opportunities (Cont’d) Ø Florida Model Applications Conference (held every other year) Ø FSUTMS Users Groups § § § 12 Central Florida Transportation Planning Group Panhandle Transportation Applications and FSUTMS Users Group Southwest Florida FSUTMS Users Group Northeast Florida Transportation Applications Forum Southeast Florida FSUTMS Users Group Tampa Bay Applications Group

New Trends in Modeling Ø Time-of-day modeling Ø Activity-based modeling Ø Integrated transportation and land use models Ø Advanced toll modeling procedures Ø Dynamic traffic assignment Ø Planning for Autonomous/Connected Vehicles 13

New Trends in Modeling (Cont’d) Ø Time-of-day modeling § Time-of-day models disaggregate the daily traffic volumes into peak periods or hours § Current daily models address capacity related issues and evaluate the impacts of new developments § They are inadequate in handling issues that require estimation of traffic variations during different time periods of a day 14

New Trends in Modeling (Cont’d) Ø Time-of-day modeling (Cont’d) § Examples of requirements include: – – – Transportation Demand Management (TDM) measures Transportation System Management (TSM) measures Transit (e. g. , New Starts Analysis) Analysis Intelligent Transportation System (ITS) Analysis Vehicle Emissions and Air Quality Analysis Traffic Congestion Management Systems Ø Data needs required to develop effective time-of-day models in Florida are currently being evaluated Ø The pre ABM SERPM 6. 0, TBRPM 7 and NERPM 5. 0 are time-ofday models Ø All the ABMs report outputs by Time-of-Day 15

New Trends in Modeling (Cont’d) Ø Activity-based modeling § Travel decisions are based on modeling the demand for activities rather than just trips – Treatment of travel as a demand derived from the desire to participate in other activities § Tours, as sequences of trips, can be modeled without modeling the underlying activity patterns (although most modern tour-based models are activity-based) 16

New Trends in Modeling (Cont’d) Ø Activity-based modeling (Cont’d) § Daily activity patterns have related travel patterns, which are expressed as tours (account for entire daily activity chain) 17

New Trends in Modeling (Cont’d) Ø Activity-based modeling (Cont’d) § Disaggregates trips at the household level instead of at the TAZ level § More demanding data requirements since model must predict travel behavior in detail including how the activities are scheduled and selected § Activity Based Models currently implemented in three regions – Northeast Florida – Southeast Florida – Tampa Bay 18

New Trends in Modeling (Cont’d) Ø Integrated transportation and land use models § Land Use Models use economic theories and simplified statistical methods to explain and estimate the location of urban land uses § Forecasts future land use changes and allocations and incorporate those changes into travel demand models § Helps determine economic and environmental impacts of land use and transportation policies § Integrated land use and transportation models can help capture the feedback loop of traditional four-step models by incorporating the impact of transportation and land use decisions on one another 19

New Trends in Modeling (Cont’d) Ø Advanced toll modeling procedures § Open road tolling – No delay at the toll plaza § Distance-based tolls – Considers tolls based on distance traveled § Discrete Tolls – Toll based on number of toll facilities crossed Ø High Occupancy Toll (HOT) lane or value-priced tolls Ø Ramp-to-ramp tolls Ø Congestion pricing 20

New Trends in Modeling (Cont’d) Ø Dynamic traffic assignment § More behaviorally sound approach to describe time-varying network and demand interaction § Routes and flow rates change during the model period based upon congested costs § Useful for evaluating: – Changes in roadway configuration – Travel demand management strategies, such as congestion pricing or peak spreading – HOT and HOV lanes – Transit vehicle behavior (shared guideway) 21

FSUTMS/VOYAGER MODELING TRAINING SERIES What is a Travel Demand Model? Forecasting and Trends Office Intermodal Systems Development Florida Department of Transportation

This Section: Ø What is a model? Ø Types of models Ø Traditional Four-Step Modeling Process § § 23 Trip Generation Trip Distribution Mode Choice Trip Assignment

What is a Model? A schematic description of a system that accounts for its known properties and may be used for further study of its characteristics 24

Planning Terms Ø Transportation planners use the term modeling and forecasting interchangeably as “Travel Demand Modeling” or “Travel Demand Forecasting” Ø A travel demand model is a mathematical description of a transportation system’s characteristics including land use, transportation network, and travelers Ø This spatial model is used to replicate existing traffic patterns and then to predict traffic volumes based on anticipated changes in the other characteristics 25

Basic Elements of a Model Ø A Transportation Demand Model § The ‘Model’ is the various equations and data which reflect travel behavior § The ‘Software’ is the platform to implement these equations 26

Types of Models Ø Types of Models by Mode § Highway Vehicle Models § Transit Models § Freight or Truck Models Ø Types of Models by Geography § Urban Area Models § Regional Planning Models § Statewide Models 27

The Four-Step Modeling Process Ø At the most basic level, the typical transportation demand model is structured around the following four sequential steps: § Trip Generation (How Many Trips? ) § Trip Distribution (Where To? ) § Mode Choice (Which Mode of Travel? ) § Trip Assignment (What Route? ) Ø Additional steps commonly include network building, transit, time-of-day, freight and reporting 28

Trip Generation – How Many Trips? Trip Generation estimates the number of trips produced by and attracted to each traffic analysis zone within the study area, based upon the socioeconomic characteristics (i. e. , population and employment) of each zone.

Trip Generation Example 1 2 3 4 FSUTMS SFDU MFDU xxx xxx Household and Population Zone Data EMPLOYMENT 1 xxx xxx 2 xxx xxx 3 xxx xxx 4 xxx xxx Employment Zone Data SPECIAL GENERATORS 2 xxx xx 3 xxx xx TRIP GENERATION TAZ 1 2 3 4 PRODS ATTRS xxxx xxxx xxxx xxxx ADJUSTED Special Generator Zone Data

Trip Purposes Ø In general, trip generation models are used to predict the number of trips within geographical subareas (zones), usually on a daily basis and for several trip purposes Ø The number of purposes may vary from one urban area to another, depending on the complexity of the model Ø The purposes are, at a minimum, divided into home-based work, home-based non-work and non-home-based trips 31

FSUTMS Trip Purposes Ø Purpose 1 – Home Based Work Ø Purpose 2 – Home Based Shop Ø Purpose 3 – Home Based Social/Recreation Person Trips Ø Purpose 4 – Home Based Other Ø Purpose 5 – Non-Home Based Ø Purpose 6 – Truck-Taxi Ø Purpose 7 – External-Internal Vehicle Trips Ø Purpose 8 – External-External Note: Some of the models have expanded the trip purposes to include airport, special tourist sites, school and freight trucks 32

Distinction Between Trips and Trip Ends Ø Trip generation models estimate trip ends by traffic analysis zone, not trips upon the travel network Ø The trip distribution model develops trips by matching together pairs from each of the two classes of trip ends: § Production – The home end of a home-based trip or the origin of a non-home-based trip § Attraction – The non-home end of a home-based trip or the destination of a non-home-based trip 33

Trip Definitions DA OP HOME D P ONE HOME-BASED WORK TRIP 1 Origin 1 Destination 1 Production 1 Attraction ONE HOME-BASED SHOP TRIP WORK OP ONE NON-HOME-BASED TRIP 1 Origin 1 Destination 1 Attraction 1 Production 1 Attraction O A DA SHOP 34 34

Basic Unit for Trip Productions & Trip Attractions Ø Household – Basic unit for home-based trip productions (dwelling units, apartments, hotel/motels, group quarters) Ø Employment – Basic unit for trip attractions (businesses, factories, plants, office buildings, shopping centers) 35

Socioeconomic Input Data Requirements Ø Socioeconomic data such as number of dwelling units, school enrollment, and number of employees are needed as input Ø Once a study area is defined, it must be divided into small geographical “units. ” These units are called traffic analysis zones, or TAZs 36

Delineation of TAZs 37

Types of Trips Internal-Internal (II) I I E External-External (EE) E Study Area Model Boundary I Internal-External (IE-EI) Note: II Trips have both ends of trip inside the study area EE Trips have both ends of trip outside the study area IE-EI Trips have one end inside and one end outside 38 E

Trip Distribution – Where To? Trip Distribution determines the number of trips between each pair of zones in the urban area based on the magnitude of each zone’s land uses and travel times between zones

Trip Distribution Example FSUTMS % PRODS TRIP DISTRIBUTION ATTRS FROM Person-Trip Table 1 2 3 4 1 xx xx T O_____ 2 3 4 xx xx xx

Overview of Trip Distribution Ø The Trip Distribution model is the process by which trips originating in one zone are distributed to the other zones in the study area Ø Trip distribution models generally link trip origins with trip destinations based on measures of accessibility and relative levels of activity at each destination zone Ø The Trip table contains the number of person-trips between each pair of traffic analysis zones (a number of models use different trip purpose stratifications) 41

Gravity Model Ø A mathematical model of trip destinations based on the premise that trips produced in any given TAZ will distribute themselves in accordance with the distance between TAZs and the attractions they offer Ø The Gravity Model parallels Newton’s Law of Gravity: Fgravity = G 42 M a. M b r 2

Destination Choice Ø Another approach to determine trip destination Ø Differs from Gravity model Uses probabilities (logit model) of a trip being attracted to a specific zone § Distance between origin and destination zone § Attractiveness of zone (based on trip attraction results) § Accessibility of zone (walkability, non-motorized infrastructure, other considerations) § Intrazonal availability of producers and attractors § Zonal Density (Population and Employment) Ø OD results are total of probabilities of trips traveling between respective zones 43

Mode Choice – Which Mode of Travel? Mode Choice calculates which trips will use the highway network and which will use the transit network. In other words, the models predict how the trips will be divided among available modes of travel

Mode Choice Example FSUTMS MODE CHOICE Person. Trip Table *Highway Only (Auto Occupancy Model) 1 2 3 4 1 xx xx T O_____ 2 3 4 xx xx xx *Vehicle-Trip Table FROM Transit Trip Table 1 2 3 4 1 xx xx T O_____ 2 3 4 xx xx xx

Overview of Mode Choice Ø Mode Choice is actually applied in two distinctively different ways: § For Highway-Only applications, Mode Choice is simply an auto occupancy model § For MPOs with transit models – Mode Choice determines the split among highway and transit modes of travel as well as auto occupancy – Typically implemented using nested logit routines 46

Trip Assignment – What Route? Trip Assignment Models predict the routes that the trips will take, accounting for diversion caused by traffic congestion. The assignment model results in traffic forecasts for the highway system and ridership forecasts for the transit system.

Trip Assignment Example Transit-Trip Table Vehicle-Trip Table HIGHWAY/TRANSIT ASSIGNMENT Network Link Volumes From To Node Volume 1001 1002 xxxx 1001 1003 xxxx 1002 1004 xxxx 48

Overview of Trip Assignment Ø What are the results of trip assignment? § Number of trips on each link in the highway network – Daily (24 hour) – Period Assignment (if model so equipped) § Number of transit riders on each route in the transit network 49

Standard Results Ø Reporting Highway Assignment § Vehicle Miles Traveled (VMT) and Vehicle Hours Traveled (VHT) § Number of trips loaded and counted § Volume/Capacity ratios by VMT, VHT, and Peak Season Weekday Average Daily Traffic (PSWADT): – By area type, facility type, and number of lanes 50

Standard Results (Cont’d) Ø Reporting Highway Assignment (Cont’d) § Free-flow and congested speeds and percent difference. § Additional for validation: – Root Mean Square Error (RMSE) by volume group (variance) – Volume/Count ratios by area type, facility type, and number of lanes 51

Standard Results Ø Reporting Transit Assignment § § § 52 Boardings and alightings by stop and direction Estimated number of transfers by mode and route Peak period transit vehicle requirements Transit station loading report in tabular and graphic format Estimated ridership by mode and route compared to observed values

Displaying Results (Cont'd) Ø Mapping § Boardings and alightings § Transit ridership flow § Transit desire lines 53

Summary Ø We covered: § Types of Models § Overview of four-step modeling process – Trip Generation – Trip Distribution – Mode Choice – Trip Assignment 54

FSUTMS/VOYAGER MODELING TRAINING SERIES Who Uses Travel Demand Models and What Are They Used For? Forecasting and Trends Office Intermodal Systems Development Florida Department of Transportation

This Section Ø In this section we will learn who uses travel demand models and how they are used in Florida Ø We will also give you some examples of model applications 56

Who Uses Travel Demand Models? Ø In this section we will look at the following to help you understand who uses travel demand models: § § 57 Users of Models Level of Expertise Required to Develop and/or Apply the Models Florida’s Distribution of the Models Exchange of User Knowledge

Users of Models Ø Metropolitan Planning Organizations (MPOs) Ø Regional Planning Councils (RPCs) Ø City and County Planners Ø FDOT District Offices Ø FDOT Central Office (Statewide Models, Regional Models) Ø University Staff (Research and Teaching) Ø In-State and Out-of-State Consultants 58

Levels of Expertise Ø Supervisor/Manager: Basic understanding of the overall modeling process and how the models are applied Ø Basic: Understanding of overall modeling process, including inputs and outputs and the ability to execute the modeling process for testing projects Ø Applier: Detailed understanding of modeling process, including the ability to maintain, apply and analyze model outputs for a wide variety of analyses Ø Developer: Detailed understanding of software and modeling process, including the ability to design, create and implement new model processes 59

Distribution of Models Ø In Florida, the MPO Models are normally developed, maintained and distributed by the MPO with support of the district office staff § In some cases the models are maintained and distributed by the FDOT District staff or by the Regional Planning Council (RPC) with assistance and support of the local MPO 60

Distribution of Models (Cont’d) Ø Once the model is adopted by the local MPO, interested planning agencies/consultants can request a copy of the model with a simple request to the distribution agency § The model data is distributed free of charge Ø The requesting agency/consultants are responsible for any changes made to the model Ø FDOT District Offices distribute any regional models by district 61

Overview of FSUTMS Software Ø FSUTMS, Florida Standard Urban Transportation Model Structure, is a computerized transportation planning model package developed by the Florida Department of Transportation and Florida’s Model Task Force Ø FSUTMS is used by all 26 Metropolitan Planning Organizations, FDOT Districts and other planning agencies in Florida Ø FSUTMS is being updated to reflect the increased presence of advanced modeling capabilities § ABMs – What aspects may be standardized? – Network attributes, socioeconomic inputs etc. ? § Freight modeling – Supply chain – Can it be integrated into FSUTMS? 62

Exchange of User Knowledge Ø The Forecasting and Trends Office offers the following forums for exchange of user knowledge: § Model Task Force (MTF): – Full MTF – Four committees – Model Advancement, Transit, GIS, Data § User Groups – FDOT supports six user groups: – – – 63 Tampa Bay Applications Group Central Florida Traffic Data Users’ Group Panhandle Transportation Applications and FSUTMS Users’ Group Northeast Florida Transportation Applications Forum Southwest Florida Users’ Group Southeast Florida Users’ Group

Exchange of User Knowledge (Cont’d) Ø The Forecasting and Trends Office offers the following forms of exchange of user knowledge (cont’d): § Quarterly email newsletter § FSUTMS Model Training Series § FSUTMS web portal 64

Uses of Travel Demand Models Ø MPO Plan Updates Ø Ø Comprehensive Plans Ø Ø SIS/FIHS Planning Ø Ø Campus Master Plans Ø Concurrency Applications Ø Ø Congestion Management Systems Ø Ø Corridor Studies Ø Ø Freight Studies Ø Air Quality Analysis ITS Evaluation Strategies Project Development & Environment (PD&E) Studies Interchange Justification /Modification Reports (IJR/IMR) Development of Regional Impacts (DRIs) Toll Feasibility Studies Economic Analysis

Example Applications of Models Ø Determine impact of specific transportation improvements Ø Determine impact of proposed developments Ø Determine origin and destination of trips Ø Determine the type of trips being made in the region (trip purpose) Ø Determine use of truck lanes, High Occupancy Vehicle lanes (HOV), and toll facilities Ø Statistical comparisons 66

Determining Impacts of Specific Developments and Transportation Improvements Ø Comparing traffic volumes § Labeling traffic volumes § Viewing higher volume roads by the thickness of their lines (multibandwidth) Ø Comparing travel times § Minimum travel times can be mapped based on shortest distance or congested speed Ø Comparing transit ridership by route and station or stop if desired (although systemwide transit ridership is likely more reliable than route-level ridership) 67

Determining Origin and Destination of Trips Ø Travel demand models can be used to determine the origin and the destination of trips (assists with determining travel patterns) Ø The user can select a specific TAZ to determine the number of trips coming from and going to each TAZ in the model (Selected Zone analysis) § Maps showing the lines to and from each zone, with thicker lines representing more trips, are known as Desire Line maps Ø The user can also select a specific roadway link to determine the number of trips and routes used for those traversing the selected link 68

Desire Line Map Example 69

Determining Use of Truck Lanes, HOV Lanes, and Toll Facilities Ø If a model includes a separate facility type or flag for exclusive truck or HOV lanes, the impacts of implementing truck or HOV lanes can be evaluated within the model Ø If a model includes a toll facilities component, toll facilities may also be tested within the model 70

Comparing Statistics Ø Comparing systemwide statistics are useful when determining the impacts of alternatives Ø Examples of statistics that may be evaluated include, but are not limited to: § Vehicle Miles Traveled (VMT) § Vehicle Hours Traveled (VHT) § Volume-to-Capacity Ratio § Congested Speed § Average Travel Times 71

Summary Ø We covered: § § § 72 Who uses travel demand models Levels of expertise Training requirements Overview of models and software Model applications

FSUTMS/VOYAGER MODELING TRAINING SERIES How do I know if I have a good model? Forecasting and Trends Office Intermodal Systems Development Florida Department of Transportation

This Section Ø In this section: § The model is just a tool § Model calibration § Model validation 74

The Model is just a Tool Ø Travel demand models are used as a tool and should not be considered the “final word” on policy decisions, etc. Ø Local knowledge and logic should also be applied when considering model results 75

Model Calibration and Validation Ø What is the difference between model calibration and model validation? § Model Calibration is a process where models are adjusted to simulate trip making characteristics of households in the model study area to match observed traffic activity in the study area; Trip rates, friction factors, and auto occupancy rates are derived from the survey § Model Validation is a process where models are adjusted to simulate new base year traffic counts; Rates and factors are derived from earlier local studies or borrowed from studies outside the area 76

Model Calibration and Validation (Cont’d) Ø When is a model calibrated and validated? § At the beginning of a long range transportation plan update or when updating to a new base year § When developing new models from scratch § After refining all or portions of a model (typically validation only) § Models are calibrated and validated to a base year to ensure the model is simulating existing conditions § Once the base year model is calibrated and validated, future year models can be developed 77

Model Calibration Ø How do you calibrate a model? § Trip Generation – Derive Trip Rates/ Trip Behavior from survey data § Trip Distribution – Derive Friction Factors, Destination Choice Logsum parameters from survey data § Mode Choice – Derive Auto Occupancy Factors from survey data – Derive Transit Ridership from survey data 78

Model Validation Ø How is a model validated? § Verify accuracy of input data and conduct quality assurance checks – Socioeconomic Data • • • 79 Conduct windshield surveys Conduct telephone sample interviews Compare with licensing databases Use telephone directories to supplement address matching efforts GIS mapping of data by TAZ Compare to County totals from the Bureau of Economic and Business Research (BEBR), U. S. Bureau of Economic Analysis (BEA), and previous models

Model Validation (Cont’d) Ø How is a model validated? § Verify accuracy of input data and conduct quality assurance checks (cont’d) – Network Data • Verify facility type, area type, and number of lanes • Verify traffic count numbers and locations • Verify centroid and centroid connector locations (TAZ centers of activity and access points) 80

Model Validation (Cont’d) Ø How is a model validated (Cont’d)? § Iteratively adjust model parameters where appropriate § Iteratively check and correct earlier data assumptions § Compare against other sources – Compare against previous datasets – Compare against datasets from comparable regions § Compare results to established accuracy standards – FHWA Model Validation and Reasonableness Checking Manual – FHWA Quick Response Freight Manual – FDOT Model Update Task C 81

Model Validation (Cont’d) Ø How do you know when the model is validated? § Output statistics to be reviewed and compared: – Trip Generation • Total trips by purpose • Aggregate trip rates: trips/household, trips/person, work trips/employee • Persons per household – Trip Distribution • Average trip length • Intrazonal trips 82

Model Validation (Cont’d) § Output statistics to be reviewed and compared (Cont’d): – Mode Choice • Mode share by trip purpose – Trip Assignment • Vehicle Miles Traveled (VMT), Vehicle Hours Traveled (VHT), and Volume-to-count ratios by facility type, area type, and number of lanes • Volume-to-count ratios by screenline • Root Mean Square Error (RMSE) by volume group • Transit ridership by route 83

How do ABMs Differ re Calibration / Validation Ø ABMs – Series of Choice Models Ø More levers to calibrate § Increased complexity § Increased data requirements Ø Each lever needs to be separately calibrated Ø At end, similar aggregate assignment metrics § § Overall Volume-over-Count ratios %RMSE Screenline comparisons Transit ridership by route etc.

Summary Ø We covered: § Model is just a tool § Model calibration § Model validation 85

FSUTMS/VOYAGER MODELING TRAINING SERIES What do I need? Forecasting and Trends Office Intermodal Systems Development Florida Department of Transportation

This Section Ø In this lesson we will introduce you to: § Hardware requirements § Data requirements § Software requirements 87

Hardware Requirements Differs based on Model type: Ø Very Simple Trip Based covering small geography and < 100 zones § Pentium 2. 8 MHz CPU § 4 GB RAM § 80 GB Hard Drive § Color Display § Windows 7, Windows 10, UNIX, Mac OS Ø Note: The larger and more complex the model, the higher the minimum requirements. 88

Hardware Requirements (Contd. ) Ø Activity / Tour Based § At least 16 effective cores § Typically Intel Core i 7 CPUs § 32 GB RAM § 500 GB Hard Drive § Color Display § Windows 7, Windows 10, UNIX, Mac OS Ø Note: The larger and more complex the model, the higher the minimum requirements. 89

Data Requirements Ø Distinction between model data needed to run the model vs. data needed to develop a model dataset 90

Data Required to Develop a Model Dataset Ø Data are available free of charge Ø Some data require additional agency cost § Depends on complexity of model § Some of the localized data may be recent enough that it still accurately depicts travel characteristics of the region § Sometimes data can be borrowed from comparable regions but will not be as accurate as local data 91

Typical Data Requirements Data Free of Charge – Included in Developed Models Data Categorized by Source Population Single-family, multi-family Hotel-motel occupants U. S. Census Visitor Data Dwelling Units Single-family and multi-family; Percent vacant and percent non-permanent U. S. Census Auto Availability Single-family and multi-family; Percent 0, 1, 2+ vehicles available per household U. S. Census Hotel-Motel Units Total hotel-motel units; Percent occupied Local Chambers of Commerce Department of Business and Professional Regulation Employment Industrial, commercial, and service employees by place-of-work Statewide Info. Group (fka Info. USA) database (2010 data) purchased by FDOT available free of charge to MPOs 92

Typical Data Requirements (Cont’d) Data Free of Charge (Cont’d) Data Categorized by Source School Enrollment School location Local School Boards Parking Cost Short-term and long-term Individual paid parking lots/garages Traffic Counts Already Available 24 -hour traffic counts for daily models and by 15 -minute intervals for time-ofday models County or City Public Works Departments and FDOT Traffic Count CD Roadway Characteristics Facility type, area type, and number of lanes for those roads included in model County or City Public Works Departments and/or road characteristic inventory and other GIS databases 93

Typical Data Requirements (Cont’d) Data Free of Charge (Cont’d) Data Categorized by Source Toll Facility Data (if toll Location, type, toll fares, service time, ratio facilities included in of heavy trucks model) Local toll authorities or Florida’s Turnpike Enterprise Transit Network Data (if transit network included in model) Local transit agencies 94 Route, stop, and station locations, mode, speed, time, fare, headway, passenger capacity, park & ride locations and capacities, short- and long-term parking cost, walk time from park & ride and kiss & ride to station

Typical Data Requirements (Cont’d) Data Requiring Additional Agency Cost Data Categorized by Source External Trips Internal-External and External by origin/destination zone pairs Origin/Destination Survey (recommended if more than 10 years old; estimated cost $100 k-200 k) Trip Production and Attraction Rates Trip production and attraction rates by trip purpose and attraction rates by employment category, total dwelling units, and school enrollment Household Travel Diary Surveys (recommended if more than 10 years old; estimated cost $100 k-300 k) Supplemental Traffic Counts 24 -hour traffic counts for daily models and by 15 -minute intervals for time-of-day models County or City Public Works Departments or contract with data collection firms (estimated cost $200 -1, 000 per location) Mode Choice Constants and Coefficients (if transit included in model) Trip purpose by 1, 2, 3+ persons per auto and transit modes Local on-board transit surveys or stated preference surveys (recommended if more than 10 years old: estimated cost $100 k$200 k) 95

NHTS Add-On ►NHTS: National Household Travel Survey § National inventory of daily and long distance travel § Only source of information at national level on the relationships between the characteristics of personal travel and demographics of the traveler § Sampling Unit – Household and Household Members § Conducted every 5 to 7 years (most recently 2016) – Results not yet available § The Metropolitan Planning Organization Advisory Council (MPOAC) purchased the 2009 add-on which included additional samples throughout the State of Florida. – 15, 884 Households – 6. 2% 96

American Community Survey Ø Replaced CTPP (Census Long Form) Ø Annual, rolling survey began in 2005 Ø 2011 -2015 5 -Year data released Jan 19, 2017 Ø Sampling Unit – Household and Household Members. Ø Sample rate ~2% Ø Greater Geographic resolution with 5 -year rolling averages – down to block group level Ø See www. census. gov/acs/www/ 97

Longitudinal Employer-Household Dynamics Ø Based on unemployment insurance reporting Ø Provides home and work locations of workers – work tours only Ø Updated regularly (quarterly for QWI data) Ø Covers 90% of civilian jobs Ø Does not include Federal employment, farm & agricultural workers, domestic workers, self-employed and armed forces Ø Available at Census Block Group Level Ø See www. census. gov/led/ 98

FSUTMS Ø Status of FSUTMS standards and impact on models § What is the purpose of having FSUTMS standards? – Ensures integrity of model structure – Portability of model from one region to another – Improves data sharing capabilities – Streamlines technical support process § What do the standards cover? – Model structure and application issues – Zonal data and network file attributes – Directory structure and file naming convention – Parameters, templates, and reports 99

FSUTMS (cont. ) Ø Status of FSUTMS standards and impact on models § § FSUTMS standards are currently under review Accommodate ABM’s and other advanced techniques What aspects can be standardized? Does it make sense to have separate standards for separate levels of model complexity? § Surveys among Florida model users undertaken – Reponses will be used to guide process 100

Summary Ø We covered: § Hardware Requirements § Data Requirements § Software Requirements 101