Modelling Data Collection and Workflow Planning

Steven Piotrowski, SMEC - Fail to Plan/Plan to Fail

Steven Piotrowski

Technical Principal | SMEC

Steven Piotrowski is a Technical Principal - Transport Modelling with SMEC based in Perth, Western Australia.  He has more than 30+ years of experience in transport modelling, public transport planning and traffic engineering.  Steven is a Fellow of Engineers, Australia and a Fellow of the AITPM. 

Fail to Plan/Plan to Fail

Clients of transport modelling are demanding more complex models that are capable of considering new modes (micro-mobility and ride hailing), new technologies (electric vehicles and autonomous vehicles) and changes in the way we travel (more working from home, more online shopping and reduced trip making per person). But without recent household travel survey data (and lots of it), our existing transport models will not be able to meet these complex challenges regardless of the complexity of our models.  

In addition, more complex models have higher run times and are more expensive to build and maintain. Models are only a tool and they are only as good as the data used to create them.

Modellers are often asked to model a narrow range of network and land use scenarios with tight timeframes and unrealistic budgets after a decision about the preferred option has already been made.  They can be used to test a wide range of scenarios but we need to allow enough time for this and we need the support of decision-makers to be allowed to examine a wider range of options.

Marius Roman, Transurban - Automation of model run, reporting and governance

Marius Roman

Principal - Transport Modelling | Transurban

Marius is a transport modeller with 15 years' experience in travel demand forecasting. He has worked on a large variety of projects for both public and private sector clients, more recently focused on the development, application and evolution of Transurban's regional strategic models.

Co-Author(s): 

Mathew McKerrow | Transurban

Mathew is Principal of Model Development at Transurban with over 20 years' experience in applying and developing strategic traffic models for a wide range of purposes.   In his current role, Mathew is responsible for maintaining, developing and applying Transurban's three regional models of Brisbane, Melbourne and Sydney to deliver high-quality robust traffic and revenue forecasts, not only for the current assets that Transurban own but for those they wish to acquire.  

Key projects\acquisitions that Mathew has been involved include the West Connex Sale Bid, Queensland Motorway Sale Bid and Westgate Tunnel project.   

Automation of model run, reporting and governance

Modelling practitioners tend to focus on innovation of the modelling tools and seldom focus on improving the framework/processes in which the models are applied. Whilst our models have developed in line with the requirements of our stakeholders, the processes in which we apply them has largely remained the same. These manual model application and reporting approaches are inefficient, unstandardised, and can omit the recording of important information for governance activities. 

The Strategic Model Automated Controller (SMAC) was developed by Transurban to solve these issues and modernise the model application framework. Showcasing the SMAC application and the advantages it brings should hopefully encourage modelling practitioners in the AITPM community to also consider the development of similar tools to assist with management of the modelling workflow. 

Following a brief introduction to the topic, the presentation will focus on four main themes, namely: 

1. Changes required to the strategic model whilst Transurban's strategic models use Cube software, the management processes described in this paper can be similarly applied in any modelling system. 

2. Front-end User Interface (UI) - The SMAC UI is essentially a Kanban board where issues are created and assigned to workflow lists.  

3. Back-end processes -  This includes processes to start, manage/monitor and close out model runs and to extract, transform and load data into readily accessible databases. 

4. Reporting dashboards, templates and monitoring tools for management and reporting of model run information and model run results. 

The presentation will end with a list of further developments that we have identified that will make the SMAC application even more indispensable to our modelling team.  

Yvonne Chen, SMEC - Requirements for HTS Data its current use in building transport models and future potentials

Yvonne Chen

Graduate Transport Modeller| SMEC

Yvonne is a Graduate Transport Modeller in SMEC's Transport Planning & Advisory team.  She has developed microsimulation and strategic modelling skills at work. Yvonne is passionate about applying data analytics skills to find efficient and innovative solutions to complex transport planning problems

Co-Author(s):

Steven Piostrowski | SMEC

Steven Piotrowski is a Technical Principal - Transport Modelling with SMEC based in Perth, Western Australia. He has a unique range of experience in transport planning, finance and property.

Steven has more than thirty years of experience in transport planning, transport modelling, public transport planning and traffic engineering in Canada, the UK and Australia.  He has also undertaken consulting assignments in New Zealand, Malaysia, Indonesia and Hong Kong. 

Steven is a Fellow of Engineers, Australia and of the Australian Institute of Traffic Planning and Management (AITPM).  He is a member of the AITPM's national transport modelling network (TMN).
 

Requirements for HTS Data its current use in building transport models and future potentials

The quality of forecasts produced by a transport model is highly dependent on the quality of data used to calibrate and validate it. Household travel surveys (HTS) are an essential data source for strategic model calibration. It is therefore crucial to understand the content of existing HTS's in Australia and potential improvements for future practice. 

This presentation demonstrates a case study conducted in 2021 where a destination choice model was developed for shopping trips in Metropolitan Melbourne based on the VISTA data. We examine characteristics of people's travel patterns for non-work trips such as travel time, distance and the number of shops and supermarkets at the destination end of the trip. We were also able to identify some gaps in the HTS including:

- Trip inclusion relating to trip chaining issues: shopping trips made on the way to home/work would not be recorded as shopping trips but as non-home based trips which was problematic.
- Trip purpose breakdown. More detailed trip purpose information may have been useful such as what type of shopping/entertainment trip rather than a general shopping trip. 
- Zones size is categorised using SA1s - statistical areas whose boundaries are decided by population alone (not employment), resulting in vastly different zone sizes in different geographic areas.
 
The presentation further discusses the direction where HTS's can be improved, including the minimum desirable sample size required and the potential and challenges of using big data to facilitate filling in the gaps.
 

David Brown, Driven Media - Why mentoring must be more than just technical excellence

David Brown

Traffic Engineer & Planner | Driven Media

David Brown has a degree with honours in civil engineering which led to a passion for traffic engineering and transport planning. He is particularly interested in how and why we use our transport systems.

Since practicing in government and private industry he has moved into the communication of technical issues not only to other professionals but to also engaging with the broader public.  He is heard on over 50 radio stations each week and his half hour radio program (which involves serious and light-hearted subjects) is also a podcast that was recently nominated as one of the top ten podcasts for transport.

He was once asked to participate in a debate about future transport not just because of his expertise but because of his ability to think broadly and inspire others to explore together what is plausible, preferred, and what needs to be changed in order to get there.

Why mentoring must be more than just technical excellence

STEM professions focus on the aspects of good scientific research methods: Data collection, analysis, hypothesis and conclusions.
As young people develop in our professions, we often become engrossed in these measures which is a good thing. Intellectual rigor and enthusiasm are worthy aspects to aspire to.

But the success or otherwise of our lives and the value they have to people around us is impacted by our attitudes and our responses to the environment around us.

For our profession to prosper (in the sense of doing good for the community not just in terms of profit), we need to heighten our skills and our support of young and old colleagues in areas other than just the specific field in which they have focused.

To emphasise this point it is worth looking at several areas where we have had major technical developments yet the impact has been distorted towards sub-optimal solutions.

Transport modelling 
From the early days of looking at just journey-to-work to more recent developments in the varying nature of many trips, modelling had progressed.

But it is still too often used in the context of giving a definitive answer at the end of a project just before, or even surprisingly after a decision has been made.

Scenario planning is a positive direction but the narrative that is created from the analysis is too often still in the hands of those with a preconceived idea of what should be done.
One major change should be to see modeling (which involves data collection, assumptions and analysis) could provide more insights during the process.

Prof Hensher's approach for Sydney University could totally revolutionise community consultation in conjunction with modelling if it is applied in a different way to what modelling is done at present.

Road Safety 
Behavioural science is, fortunately, an expanding part of road safety yet our impact is stultified by an approach of - "we have found the answer so we have to tell everyone so that they will change." This lecture rather than engagement approach is extremely limiting. 

This weakness is continually being shown in some professional presentations, government advertising and some areas of driver training. Working with some behavioral scientists, we presented some papers on a new approach. It produced positive responses in a better way to approach problems in general not just road safety. It showed some underlying philosophies that must be part of the skills we acquire to get the best results.

Human Machine Interface 
The advent of digital technology has greatly increased the amount of information and the way it can be displayed in cars.

But the human/machine interface must not be driven by a desire to create greater use of technology to the point where the result is more distraction rather than helpful information.

We all have an opinion about what we like and what we are familiar with. But this can lead to heated debate but non-productive results.
We are working with Prof Mike Regan and some other significant people in the field of ergonomics to establish ways to measure the sort of distractions that are happening with modern vehicle interfaces.

The key step is to abandon the approach of working away in isolation to try and produce the perfect result. We are looking at ways to identify what is really happening and working with stakeholders to instill proper procedures that can adapt to future developments.