LIDA Societies x DSDP Seminar Series: Healthy Lifestyles & Consumerism
LIDA: Societies have teamed up with the Data Scientist Development Programme to bring you a seminar series full of diverse topics within the field of Societies.
Thursday 9th February - 12noon -1pm
Hybrid – 11.09 / online
Speakers: Tamara Garcia del Toro, James Battye, Modupe Aggreh and Rayan Onyonka
Talk 1: Consumer Trends in Take Away Purchasing
Speaker: Tamara Garcia del Toro
Purpose:
Current research in people’s diet habits has been very focused in the food environment: the different contexts in which people engage with the food system. Originally, this concept referred to the physical presence of food in a person’s surroundings, which affects their ability to access different foods.
The food environment has been transformed in the past decade, with the development of new services such as online grocery and take away delivery services. Alongside a shift towards more out-of-home-food consumption and the unique current historical context (COVID-19 pandemic and the cost-of-living crisis).
To better understand consumer habits around take-away purchasing, and how the growth of online food delivery services has shaped new behaviours, we have partnered with a large online take away delivery platform to use their transaction data in order to shed light on how changing customer habits are shaping the food environment.
Methods:
Transaction data for online food purchasing was provided by the data partner, a large online food delivery service. The data included anonymised customer reference id. Data was accessed through the retailer’s own secure platforms. Data analysis was carried out in two phases: different machine learning methods were evaluated and used to produce a classification of customers based on purchase frequency in various categories of food purchases; and an exploration of the locational characteristics of these classifications and distribution across UK geography.
Spatial clusters of these classifications of purchasing habits across the UK were characterised using socio-demographic data, matched by consumer postcode information through geospatial analysis. Geodemographic data was sourced from the 2011 and 2021 census at the Output Area Level (Approximately 125 households) and retailer’s data was matched using postcode information.
Conclusion:
Studies to date have shown that areas with higher levels of deprivation have higher access to take-away food outlets, but it remains to be seen whether this higher access translates to higher purchasing in this communities. We are the first team to investigate customer transaction data to go beyond customer access and see what the actual online food ordering trends are.
Talk 2: Towards a Digital Twin of a Supermarket Loyalty Scheme
Speaker: James Battye
Abstract:
Digital twins represent a virtual analogue of a real-world system or object. The virtual twin is informed by its real-world counterpart through a flow of data between the real and virtual twins, potentially in real-time. Digital twins facilitate the asking of questions which may otherwise be unfeasible to test in the real world due to time, resource or ethical constraints. Questions asked of digital twins have more commonly involved understanding physical components, such as in a manufacturing context. However, more complex twins are now being developed which incorporate individual agents. This enables the trialling of ‘what-if’ scenarios, whereby an intervention can be tested utilising the virtual twin to observe agents’ behaviour change in response to the intervention.
This project is in collaboration with a large supermarket chain and aims to take steps towards the implementation of a digital twin to simulate customer behaviour change in response to a loyalty scheme. A digital twin in this context would facilitate the testing of ‘what-if’ scenarios, for example the introduction of products to the scheme, and their impact on customer behaviour. The twin will be developed using transaction and demographic data provided by the partner. Using this data, key elements and metrics necessary to drive the digital twin will be identified. This will form the basis from which a synthetic population will be created which incorporates individual behaviours regarding propensity to purchase certain goods. Research will be conducted within a secure cloud-based environment and new cloud-native digital twin-focused tools will be explored to conduct simulations.
Talk 3: FIO-FOOD, Food Insecurity in people living with Obesity – improving sustainable and healthier food choices in the retail FOOD environment
Speaker: Modupe Ahhreh
Background:
The relationship between food insecurity, development of obesity and how effective interventions are in promoting healthy eating is unclear. This project analyses both retailer data and sustainability metrics across three environmental indicators (greenhouse gas emissions (GHGE), biodiversity loss and water use) to determine the sustainability footprint of the food and drink products shoppers buy. Additionally, the project aims to examine how the pattern of food and drink purchases are linked to the Eatwell guide which is beneficial for a healthy and more sustainable diet thereby improves sustainable and healthier food choices in the retail industry.
Methods:
A product dataset which comprised of >25,000 unique products and contained nutrition as well as weight information was obtained from Sainsbury’s supermarket. Using published data, estimates of GHGE, biodiversity loss and water use were paired to the products. The combined output was appended to an anonymised Sainsbury’s loyalty card transaction dataset comprising of regular customers located in the Yorkshire and Humber region for the period of January 2016 to December 2020. The transaction dataset also includes Gender (if provided), Age Band, Area Level Deprivation as well as the Output Area Classification. Within the transaction data, products are mapped to the food groups in the Eatwell Guide. The final output will be merged into a tool for interactive visualisation purposes for Sainsbury’s.
Talk 4: Evaluating strategies to promote healthier and more sustainable dietary choices
Speaker: Rayan Onyonka
Abstract:
Poor diet is a leading cause of mortality in the United Kingdom, with one in seven deaths attributed to it. The UK’s food-based dietary guidelines (the Eatwell Guide) set recommendations for the proportions of each food group required for a healthy balanced diet. Yet, just 1% of the population currently meets these recommendations. Recent research shows that population dietary shifts towards the Eatwell Guide would be healthier and more sustainable.
The study assesses the impact of in-store promotions in supermarkets on encouraging consumers to swap their usual products for healthier alternatives and if promoting these "healthier swaps" led to positive shifts in the baskets towards the Eatwell Guide.
The trial was a non-randomised controlled before and after design conducted in all the retailer’s stores within the UK. This study limits itself to two regions: Yorkshire and the Humber and the South East. Eight own-brand products were selected based on their nutritional benefits and were the same price or cheaper than their "less healthy" equivalent. Store-level daily sales (units, weights (g) and spend (£)) were used in primary analyses.
The trial took place in February 2021 and ran for four weeks. For analysis, we received data for the 12-week pre-trial period in 2020-2021 and post-trial period in 2021. We also had two years of control data from 2019-2020 and 2018-2019 that were used to observe the purchasing history of the items before and during the pandemic.
The efficacy of the intervention and its effects were analysed using interrupted time series (ITS). Text matching algorithms were used to assign the items to their respective Eatwell category, and the basket-level shifts were tracked to observe if the trial influenced purchasing habits.