Understanding mobility patterns: Difference between revisions

Large amounts of geolocated traces are being collected that enable the analysis of mobility patterns. This can be useful for better managing existing as well as planning future transport systems. ML offers great potential to progress the following areas.

To better understand existing transport systems, '''machine learning can be used to detect the structure of a transport network as well as its utilisation'''. In the context of highly congested cities (especially with parked cars), this could, for example, support the discussion for a more equal street space allocation and the adoption of more low carbon mobility, such as walking or riding bicycles<ref>{{Cite journal|last=Creutzig|first=Felix|last2=Javaid|first2=Aneeque|last3=Soomauroo|first3=Zakia|last4=Lohrey|first4=Steffen|last5=Milojevic-Dupont|first5=Nikola|last6=Ramakrishnan|first6=Anjali|last7=Sethi|first7=Mahendra|last8=Liu|first8=Lijing|last9=Niamir|first9=Leila|last10=d’Amour|first10=Christopher Bren|last11=Weddige|first11=Ulf|date=2020-11-01|title=Fair street space allocation: ethical principles and empirical insights|url=https://doi.org/10.1080/01441647.2020.1762795|journal=Transport Reviews|volume=40|issue=6|pages=711–733|doi=10.1080/01441647.2020.1762795|issn=0144-1647}}</ref>.

Furthermore, the '''analysis and classification of travel patterns''' serves to better understand urban mobility flows. As human mobility in urban areas follow strong regularities<ref>{{Cite journal|last=González|first=Marta C.|last2=Hidalgo|first2=César A.|last3=Barabási|first3=Albert-László|date=2008|title=Understanding individual human mobility patterns|url=https://www.nature.com/articles/nature06958|journal=Nature|language=en|volume=453|issue=7196|pages=779–782|doi=10.1038/nature06958|issn=1476-4687|via=}}</ref>, this understanding constitutes the basis for making short- and long-term predictions of mobility demand.

Finally, machine Learning can be used to assess how a given '''local context, such as the urban form, impacts mobility'''. Building up on this, planning strategies can be derived which support a low carbon transport future.

'''“Urban Human Mobility Data Mining (2016)”'''<ref>{{Cite journal|last=Zhao|first=Kai|last2=Tarkoma|first2=Sasu|last3=Liu|first3=Siyuan|last4=Vo|first4=Huy|date=2016|title=Urban human mobility data mining: An overview|url=http://ieeexplore.ieee.org/document/7840811/|journal=2016 IEEE International Conference on Big Data (Big Data)|location=Washington DC,USA|publisher=IEEE|volume=|pages=1911–1920|doi=10.1109/BigData.2016.7840811|isbn=978-1-4673-9005-7|via=}}</ref> ''':''' An Overview This paper provides a great overview of the complete modelling process, from finding potential data sets, cleaning and preprocessing data sets, searching for patterns using ML, evaluating models and obtained patterns as well as acting on discovered knowledge. When it comes to finding mobility patterns, the authors also highlight different predictors and their applicability in different contexts.