CASE STUDY: Tesco Grocery Dataset 1.0 vs Income
The following case study aims to understand and analyse the Tesco Grocery Dataset 1.0 [1] which contains a large dataset of food purchase records of fidelity card owners of Tesco over the year 2015 in the Greater London region. The study also attempts to examine income data for the same year and explore the correlation between income and food purchase patterns.
Assumptions and biases:
• The dataset contains data recorded for purchases from Clubcard owners of Tesco which is not a representation of the entire population demographics. [2]
• The food purchased only captures offline purchasing habits of a particular demographic from their select choice of supermarket. Food consumption trends from other resources and supermarkets cannot be represented from this dataset. [2]
• Tesco store penetration is not uniform in all regions mentioned in this dataset. [2]
• The geographical boundary data used for plotting in this study corresponds to 2011. Any changes to boundaries beyond that are not taken into consideration.
• The food purchase records are averaged over selected regions and do not represent food consumption patterns at an individual scale. [2]
• The income datasets used only records of total average household incomes at selected geographical scales and do not represent individual household income.
Summary:
The analysis identifies a relationship between obesity and food purchasing patterns. For example, regions with high prevalence of obesity among adults had a low purchase rate of healthful foods like fruits and vegetables. The study also found that regions with a higher proportion of healthy weight adults usually had low purchase rates for sugar-sweetened drinks.
The dataset includes a total of 17 different food products that were recorded during the study. These products can be clustered into four different groups based on the total sales registered.
Product Cluster Total yearly Sales (mean) Food items Bestsellers 1331 Fruits and vegetables Popular Products 761 Sweets, Grains Standard Sales 402 Readymade and dairy items Limited Sales 97 Beverages and animal meat products
The income data for London MSOA (year 2015) [3] can be classified into four different economic classes based on mean annual household incomes. The minimum mean income is observed to be around £ 34,000 and the highest mean income was £ 84,000. Nutrition quality of food purchases varies by household incomes Socioeconomically disadvantaged groups usually prefer energy rich and nutrient deficit foods which in turn relates to tendency of obesity and overweight among adults [4,5,6]. The present study found correlations to support this result. A higher percentage of overweight adults were found to be from regions which fall in the lower economic groups from above. Households with mean annual income of £ 34,000 are about 5% less likely to purchase fruits than the highest economic class and 15% more likely to get sugar sweetened drinks like cold drinks.
References:
[1] Aiello, Luca Maria; Schifanella, Rossano; Quercia, Daniele; Del Prete, Lucia (2020): Tesco Grocery 1.0. figshare. Collection. https://doi.org/10.6084/m9.figshare.c.4769354.v2
[2] Aiello, L.M., Quercia, D., Schifanella, R. et al. Tesco Grocery 1.0, a large-scale dataset of grocery purchases in London. Sci Data 7, 57 (2020). https://doi.org/10.1038/s41597-020-0397-7
[4] d'Angelo, Camilla, Emily Ryen Gloinson, Alizon Draper, and Susan Guthrie, Food consumption in the UK: Trends, attitudes and drivers. Santa Monica, CA: RAND Corporation, 2020.
[5] Giskes, K., Avendaňo, M., Brug, J. and Kunst, A.E. (2010), A systematic review of studies on socioeconomic inequalities in dietary intakes associated with weight gain and overweight/obesity conducted among European adults. Obesity Reviews, 11: 413-429. https://doi.org/10.1111/j.1467-789X.2009.00658.x
[6] French SA, Wall M, Mitchell NR. Household income differences in food sources and food items purchased. Int J Behav Nutr Phys Act. 2010 Oct 26;7:77. doi: 10.1186/1479-5868-7-77. PMID: 20977738; PMCID: PMC2988056. [7] French, S.A., Tangney, C.C., Crane, M.M. et al. Nutrition quality of food purchases varies by household income: the SHoPPER study. BMC Public Health 19, 231 (2019). https://doi.org/10.1186/s12889-019-6546-2 [8] https://data.london.gov.uk/dataset/statistical-gis-boundary-files-london