Skills in data analysis using Geographical Information Systems (GIS) can be essential to making sense of large volumes of location-linked information for commercial decision-making.
Geographers working for Sainsbury’s apply statistical analysis and GIS to sales data to understand how shopping habits have different seasonal patterns in different places. This can then help them decide whether a place is a good location to open a new store.
In this webinar for Esri, Tim Rains CGeog(GIS) explores the seasonality of sales in UK stores. It starts with collecting two years of transaction data – trading patterns across around 1400 stores in the UK – which is then subjected to seasonal decomposition. In Tim’s examples, three patterns of trading are drawn out:
-
Underlying trends, reflecting macro factors like inflation
-
Random week-to-week variations, for example, when a competing store runs a promotion
-
Seasonal patterns – regular/recurring peaks and trading, such as increased spending before Christmas or increased sales in Cornwall in the summer, when tourists tend to visit the area in large numbers
Cluster analysis is used to combine stores into similar groups based on the seasonality of their trading, whicih produced groups including
-
Student patterns, where students were a large population – increased spending in term time (e.g. Oxford)
-
Summer stores, where spending increased in the summer holidays (e.g. Lake District)
-
Workplace stores, where trading reflected working hours or shifts of those using the shop (e.g. London and Aberdeen)
Tim also explains how predictive modelling, based on geodemographic and consumption data, can help the company identify good places to open new stores.
Link with timecodes
View the webinar here (registration required; starts 30:47)