Mining Shopping Patterns for Divergent Urban Regions by Incorporating Mobility Data

Abstract

What people buy is an important aspect or view of lifestyles. Studying people's shopping patterns in different urban regions can not only provide valuable information for various commercial opportunities, but also enable a better understanding about urban infrastructure and urban lifestyle. In this paper, we aim to predict citywide shopping patterns. This is a challenging task due to the sparsity of the available data -over 60% of the city regions are unknown for their shopping records. To address this problem, we incorporate another important view of human lifestyles, namely mobility patterns. With information on "where people go", we infer "what people buy". Moreover, to model the relations between regions, we exploit spatial interactions in our method. To that end, Collective Matrix Factorization (CMF) with an interaction regularization model is applied to fuse the data from multiple views or sources. Our experimental results have shown that our model outperforms the baseline methods on two standard metrics. Our prediction results on multiple shopping patterns reveal the divergent demands in different urban regions, and thus reflect key functional characteristics of a city. Furthermore, we are able to extract the connection between the two views of lifestyles, and achieve a better or novel understanding of urban lifestyles.

Mining urban lifestyles: urban computing, human behavior and recommender systems

“…These studies propose models for either mobility or spending behavior, but not in conjunction.The only known paper that incorporates both aspects 16 frames its analysis only on an aggregate scale of city regions. However, the coupled collaborative filtering methods (also known as collective matrix factorization) used in 16 have been successfully applied in a variety of urban computing applications for data fusion and prediction [17][18][19] , from location-based activity recommendations 20,21 to travel speed estimation on road segments 22 . Recent work includes methods that use Laplacian regularization 23 to leverage social network information, and use geometric deep learning matrix completion methods to model nonlinearities 24 .…”

Venue2Vec: An Efficient Embedding Model for Fine-Grained User Location Prediction in Geo-Social Networks

“…For example, with regard to the above movie theater in Tokyo (Japan), we will add the category vector Art & Entertainment to the location vector 4b558a35f 964a520eae627e3. Note that the extra category vector has the same dimension as the location vector 2 . In this way, locations of the same type and those are geographically near, or often visited successively by users will be closer in the shared embedding space.…”

“…This application scenario has value for both individuals and the wider society. From the individual point of view, accurate location prediction can provide users with informative personalized product recommendation [1,2]. From a societal view, such analysis can accurately predict where traffic jams would happen, thus can be helpful for urban intelligent transportation [3,4].…”

Exploring Passengers’ Dependency Variety on Stations’ Functions in Urban Subway

“…But until now, no connection has been built to connect this parameter with passengers' travel dependency [9,10], with an aim to figure out passengers' potential aims of travel. Apart from the above studies, some more research gave specific attention to characterizing mobility patterns for certain special groups, such as tourists [2], people who go shopping [11], the elderly [12], pickpockets [13,14], or passengers with extreme travel patterns [15]. However, they did not quantify their travel dependency on urban subway.…”