What If Big Data Is Not Available or Not Satisfying? a Practical Case in Parking Supply Data in Paris Metropolitan Area

Big data refers to datasets which are too big to be dealt with by traditional software. It is particularly useful in transport because it ca give a better understanding of mobility (individual and freight transport demand, user profiles…) which can help stakeholders to plan transport infrastructures and to optimise transport system. Big data can be public (e.g. public transport supply in open source) or not, depending on different issues, as commercial interests or individual privacy. Big data is increasingly used in transport modelling. Indeed, transport models are more and more powerful and complex, which explains their always increasing data requirements. These data are needed to estimate model parameters, but also for model validation and application. Some model elements where big data can be used are transport demand, public transport supply, parking supply, car speeds… But, what happens when big data is not able to answer modellers’ needs? Indeed, sometimes data is not public, and sometimes it just simply doesn’t exist. In that case, modellers need to make without these data, or get along with it and seek for other solutions. In this paper the authors will discuss an actual case the authors have been confronted with: the lack of parking supply data in Paris metropolitan area which satisfies to the requirements of the disaggregated choice model of RATP, the IMPACT model. RATP is the main public transport operator in Paris metropolitan area. In order to answer its needs in traffic modelling, RATP has two traffic models: GLOBAL, which is used mainly to simulate traffic demand in new rail lines, and IMPACT, which is used to simulate the effect of transport policies in mode and destination choices. Some examples of transport policies which can be simulated with IMPACT are a new public transport network, access restriction to cars in the central business district, new fuel costs, a change in public transport fare system… and parking supply and costs. At first, the authors tried to build this database using big data and open data. However, the authors did only find some data, for example, about on-street parking supply in Paris (only 18% of population in Paris metropolitan area) and park-and-ride data at the metropolitan level. But it wasn’t enough for us, because the authors needed all parking supply in Paris area, segmented by type and zone, and its costs. As big data wouldn’t give the authors all the answers the authors needed, the authors had to search for other solutions. This task was simplified by the fact that the authors modelling approach let the authors do some simplifications, while having enough accuracy for the model needs. In this paper the authors discuss some of the solutions the authors retained to deal with big data lack of data, which allowed the authors to build the complete parking data supply the authors needed to develop the new version of the model.

Language

  • English

Media Info

  • Media Type: Digital/other
  • Pagination: 13p
  • Monograph Title: European Transport Conference 2019

Subject/Index Terms

Filing Info

  • Accession Number: 01751302
  • Record Type: Publication
  • Files: TRIS
  • Created Date: Aug 21 2020 1:02PM