What if it was easier to predict how a city could have clean air, broader economy, increase safety and reduce traffic? The transportation engineer Josie Kressner assessed the viability of using targeted marketing data for demographic data in travel demand modeling in her dissertation, employing alternative methods for updating the transportation planning process.
City planners and engineers built what are called travel demand models to help achieve this goals, but outdated information make effective planning with these models difficult. Nationally, cities spend millions of dollars each year collecting input for these city models. The backbone data set they collected is called household travel survey. This survey obtains data about how people in the household travel in a typical day. The problem we have is this: household travel surveys are expensive! So much so that often we only collect them once every 10 years. They also suffer from large nonresponse bias and often from sampling bias too. We need to find a different and more affordable data sources to, with success, plan for the future. So how do we do that? Believe it or not, there’s a fairly simple solution right under our nose!
We live in a world where billions species of data are collected every day. These two types of data are particularly interesting for our urban planning dilemma: consumer data and anonymous mobile phone data. Both of these data are readily available and cheap. Separately, these two types of data cannot be used to replace traditional household travel surveys. In consumer data: there’s a large amount of information about people, but nothing about their trip making behavior. And in anonymous mobile phone data, there’s an actual picture of travel at any given region at any given time, but we do not know who are making these trips. However, if we combine consumer data and anonymous mobile phone data to create realistic synthetic household travel data, using statistical simulation and data fusion techniques, we can overcome their shortcomings while still respecting individual privacy. If we combine these data systematically every month, quarter or year, we will have a more accurate, more up-to-date and significantly cheaper data to help our cities get where they want to go in the future.