There’s an old saying about forecasting being like driving a car blindfolded helped by directions from a person looking out the back window. It’s well and truly time for forecasters to take the blindfold off.



Like the artist, the modeller ‘paints pictures’, but of transport conditions in the future through the medium of transport models. Rather than canvas, brushes and paint the modeller uses computers, data and mathematics.  In the same way that an artist assembles the paints of different colours into the right areas to produce something that looks like the subject, the modeller arranges matrices and algorithms and mathematical formulae to paint a picture that represents future transport conditions. Our models are our oracles – they give us bits and pieces of information about the future, but without deep understanding and human interpretation they are meaningless jumble of outputs and numbers.

While transport models contain a great deal of science, mathematics and skill in calibration and validation, the modeller needs to be an artist to design the architecture that contains the science and to try to minimise the uncertainties around unforeseen changes arising from new technologies and changes in the way people choose to travel.

We know that our city ecosystems are undergoing their biggest transformation since the invention of the car. The complexity of this transformation is augmented by rapid urbanisation, the rise of digital and disruptive technologies, population growth, changing attitudes and changing work practices, all of which are contributing to changes in the way people engage with, and move around cities that current modelling practices can’t forecast. The modeller, in many instances, is still driving with the blindfold on.

The modeller has to find a way to make models less dependent on historical trends and be more valuable for today’s decision makers faced with an unprecedented set of decisions required for a future with greater uncertainty.

When the blindfold is removed, there are a lot of opportunities to improve our models – more data, faster computers and new ideas. The modeller needs to divorce from simple trends, moving from projecting current trends to consulting with futurists to define future scenarios that represent the radical shifts we are beginning to see all around us.

The modeller needs to understand the relationships which drive behaviours. To further enhance the value of modelling, the modeller needs to not only explain ‘why something happened?’ and ‘what might happen?’ but more importantly have the model inform ‘how we can make it happen?’

In this complex landscape, policy-makers are seeking to inform their decisions with data-driven insight and solutions. Governments are working to improve transport infrastructure and customer experience against a backdrop of increasing population growth, congestion and public transport crowding. Added to this, the rise of the sharing economy, Mobility as a Service (MaaS) and autonomous vehicles offer many opportunities as well as threats introducing yet more uncertainty.

Models are works of art and for them to be useful and achieve a higher value during times of exponential change, we need to remember that models don’t forecast, humans do. And this is why VLC has launched its Future Transport Lab which is committed to removing the blindfold, working as artists and as modellers to interpret the ramblings of the oracle to know ‘how can we make it happen?’


Stay tuned for future instalments from our Future Transport Lab where we will share our views on the future of mobility and how we can use our models and big data to gather valuable insights for planning and decision-making.

This piece has been adapted from a presentation Tony Fransos made at the 2018 AITPM Conference in Perth.