[Paper Review] Correlations and dynamics of consumption patterns in social-economic networks

Yannick Léo, Márton Karsai+2 more|arXiv (Cornell University)|Jan 23, 2018

Opinion Dynamics and Social Influence63 references18 citations

TL;DR

This study analyzes a large-scale, multi-source dataset combining mobile phone communications and bank transactions from a Latin American country to link consumption patterns with socioeconomic status, social network structure, and demographic factors. It reveals strong correlations between purchasing behavior and socioeconomic class, identifies community structures in merchant category correlations, and uncovers dynamic weekly spending patterns tied to individual demographics and social ties.

ABSTRACT

We analyse a coupled dataset collecting the mobile phone communications and bank transactions history of a large number of individuals living in a Latin American country. After mapping the social structure and introducing indicators of socioeconomic status, demographic features, and purchasing habits of individuals we show that typical consumption patterns are strongly correlated with identified socioeconomic classes leading to patterns of stratification in the social structure. In addition we measure correlations between merchant categories and introduce a correlation network, which emerges with a meaningful community structure. We detect multivariate relations between merchant categories and show correlations in purchasing habits of individuals. Finally, by analysing individual consumption histories, we detect dynamical patterns in purchase behaviour and their correlations with the socioeconomic status, demographic characters and the egocentric social network of individuals. Our work provides novel and detailed insight into the relations between social and consuming behaviour with potential applications in resource allocation, marketing, and recommendation system design.

Motivation & Objective

• To investigate how socioeconomic status (SES) shapes individual consumption patterns and social network structures.
• To identify correlations between frequently co-purchased merchant categories and detect community structures in consumption networks.
• To analyze the temporal dynamics of individual purchasing behavior and its links to demographics, SES, and egocentric social networks.
• To provide empirical evidence on how social influence and economic capacity jointly shape consumption behavior at scale.

Proposed method

• Constructed a social network from mobile phone communication data to map interpersonal ties and identify ego-centric network structures.
• Assigned socioeconomic status (SES) indicators using income, occupation, education, and demographic data derived from transaction and communication patterns.
• Mapped merchant categories using standard MCC (Merchant Category Code) taxonomy from American Express, covering 271 distinct categories.
• Built a correlation network between merchant categories based on co-occurrence of transactions across individuals to detect commonly purchased goods and services.
• Analyzed weekly purchase dynamics using time-series decomposition to detect periodic patterns linked to occupation and lifestyle.
• Applied multivariate statistical analysis to correlate individual consumption dynamics with SES, gender, age, and social network features.

Experimental results

Research questions

• RQ1 Can typical consumption patterns be associated with individuals from different socioeconomic classes, and how do they vary across classes?
• RQ2 What are the underlying correlations between merchant categories, and do they form meaningful community structures in a consumption correlation network?
• RQ3 What dynamical patterns emerge in individual purchase behavior, and how are they correlated with socioeconomic status, demographic traits, and social network structure?

Key findings

• Consumption patterns are strongly correlated with socioeconomic status, with distinct spending behaviors observed across different SES classes.
• A correlation network between merchant categories revealed a meaningful community structure, indicating clusters of frequently co-purchased goods and services.
• Individuals in higher SES groups exhibit more diversified and higher-value spending patterns compared to lower SES groups.
• Weekly purchase dynamics show significant periodicity, with patterns linked to occupation and lifestyle, particularly for working-age individuals.
• Social network structure, especially homophily in SES, amplifies similarities in consumption behavior among connected individuals.
• Demographic factors such as age and gender significantly influence both the timing and type of purchases, with distinct patterns emerging across groups.