چكيده به لاتين
Over the past two decades, some types of big data, such as GPS data, have been used extensively as alternative or complementary data in transportation studies. On the contrary, the analysts have been pretty careful when it comes to mobile phone cellular data. In addition to the technical difficulties on the way, there always have been deep concerns about the users’ privacy. In this study, using anonymous mobile phone cellular data, we extracted trip production, attraction, distribution, and other travel demand characteristics in 30-minute intervals for the coverage areas of towers. Due to the fact that the data is big data, different methods of big data analytics, clustering and visualization have been used in this research. In order to validate the results, a case study has been conducted on one of the major shopping centers in Tehran. Comparing the results with the survey and questionnaire data shows that they are pretty promising in most of trip components, such as 30-minute production and attraction patterns, types of trips, distribution of trip origins, returning to the origin, and trip frequency. The trips attracted to/produced from the chosen land use, demonstrate promising correlation in 30-minute intervals (more than 0.99). In addition, the duration of people's presence, origin of travel, type of origins, and returning behavior of their travel are also largely consistent with the results obtained from mobile phone cellular data analysis. In this study, it was shown that mobile phone cellular data can be used as an alternative to traditional methods, which are based on regression and are probabilistic. The results could help many aspects of travel planning, one of which is estimating trip production and attraction in traffic impact studies.
