Spatiotemporal mapping of urban trade and shopping patterns: A geospatial big data approach

The economic viability of an urban area in terms of trade and shopping significantly impacts its residents’ quality of life and is crucial for any sustainable development initiative. Geographic information systems (GIS) are well established, but the use of GIS technology within finance and trade analysis is still in its infancy. In this article, we highlighted the potential of GIS technology and big data analytics and demonstrated the importance of thinking in spatial terms for analysing patterns within the trade and finance industries. We studied spatiotemporal trade and shopping patterns in the city of Tabriz using data generated by customer purchase transactions obtained from 5200 stores, shopping, business and service centres. We employed time series transaction data collected from the points of sale in stores, shopping, service and business centres located in different areas of the city. We applied four well known geospatial big data driven approaches including machine learning nearest neighbour, kernel density estimation, space–time pattern mining and spatiotemporal coupling tele-coupling for detecting and mapping of spatial trade hotspot patterns. The results of this study indicated the potential of GIScience methods for the explicit spatial mapping of trade and shopping patterns. The results revealed that the city centre, particularly the Bazaar of Tabriz, acts as the city’s heart of trade, and we identify additional major business hotspots. Furthermore, the results allow for studying the impacts of unbalanced urban development in Tabriz, where the wealthy suburbs with high quality of life, such as Valiasr and Elguli, host the major shopping hotspots. The spatial patterns obtained enable local stakeholders, decision makers and authorities to develop strategic plans for urban sustainable development in Tabriz. The geospatial big data approach used can stimulate novel and progressive research. Results of this study demonstrate methodological advancements in GIScience by ’spatializing’ individual purchase data and therefor proposing an explicit geospatial big data analysis approach.