This chapter deals with data production within the paradox of big-no data in urban logistics. More precisely, the chapter aims to make an overview on data production in urban logistics and present the main issues, as well as a framework, to overcome that paradox. First, the big-no data paradox is defined and motivated. Second, the question of demand and route estimation in urban logistics is examined more in-depth via an overview of current data production methods and techniques used to estimate demand and transport flows. Third, a framework to produce unified databases filling those data lacks is introduced as well as an analysis on how the different data production techniques can be used to overcome that paradox. Finally, research directions regarding urban goods transport data production are provided.
TopIntroduction
The freight transport sector, according to Crainic (2008), is faced to many issues and challenges derived from the fact that it is an essential component of a country’s economy but is a source of nuisance to the environment and the quality of life, mainly in urban areas. For that reason, the research on the field of city/urban logistics has been developed and organized in the last forty years (i.e. since Demetsky, 1974 and Watson, 1975), and started mainly by works on freight transport flow estimation. Moreover, urban freight transport characterization play a central role in many research and practice issues (Gonzalez-Feliu, 2018).
One of the main issues related to urban logistics is that of data production, since it has traditionally been difficult to produce the required data to characterize urban goods flows. Indeed, works on urban goods data collection and modelling have been for more than 40 years a heterogeneous set of methods and techniques that were mainly concluding in the impossibility to define standards, since surveys and methods were context-dependent and data little available (Watson, 1975; Ogden, 1992; Ambrosini & Routhier, 2004; Ambrosini et al., 2008; Gonzalez-Feliu & Routhier, 2012; Browne & Goodchild, 2013; Gonzalez-Feliu et al., 2013a).
With the development of new techniques, the idea that the gaps between data production and modelling seem to decrease (Gonzalez-Feliu et al., 2016). However, the issues of data production remain of actuality and although some standards start to be defined (Holguin-Veras & Jaller, 2014; Holguin-Veras, 2016), the inclusion of big data and automatic data collection techniques remain limited in urban goods transport (Allen et al., 2012; Gonzalez-Feliu et al., 2013a; Holguin-Veras & Jaller, 2014; Gonzalez-Feliu, 2018).
This chapter deals with data production in a paradoxal context: that of urban goods transport data production. The paradox is that in the big data era, the urban logistics field presents also big quantities and a huge diversity of freight transport and flow data but, on another hand, several basic informations for demand estimation and forecasting are difficult to obtain. Researchers and practitioners are confronted to this dual situation: a context mixing big data and no data situations. This no data situation is related to both a lack in capturing essential information and the confidentiality of many data sources. In this context, new methods are needed to produce unified and accessible data.
This chapter aims then to propose, via an overview of data production methods in urban logistics, a first vision of the main data issues and lacks in that field, as well as to propose a methodological framework to deal with them. First, an overview on data production in urban logistics is provided. Second, the question of demand and route estimation is examined more in-depth. Third, a framework to produce unified databases filling those data lacks is introduced, the main data needs and sources to obtain them presented and main research directions announced.