Mapping cigarettes similarities using cluster analysis methods

Area of Science:

• Tobacco product analysis
• Public health research
• Data science in consumer product evaluation

Background:

• Understanding the chemical composition, market positioning, and public health information of cigarette brands is crucial.
• Previous research has explored individual aspects of cigarette products, but a comprehensive analysis integrating these factors is less common.

Purpose of the Study:

• To investigate relationships between chemical composition (tar, nicotine, carbon monoxide), market data (brand, manufacturer, price), and public health information (class, health warning) for cigarette brands.
• To apply clustering techniques to identify groupings of cigarette brands with shared characteristics based on collected variables.

Main Methods:

• Analysis of 30 cigarette brands, collecting data on six categorical variables (brand, manufacturer, health warnings, class) and four continuous variables (tar, nicotine, carbon monoxide, price).
• Application of multiple linear regression to explore relationships between variables.
• Utilized two distinct clusterization techniques to group similar cigarette brands.

Main Results:

• A significant link was identified between carbon monoxide concentration and both tar and nicotine concentrations.
• Cluster analysis successfully grouped cigarette brands, with tar and carbon monoxide concentrations being primary differentiating factors.
• Identified distinct clusters of cigarette brands exhibiting similar chemical profiles and market attributes.

Conclusions:

• The chemical composition of cigarettes, particularly tar and carbon monoxide, plays a key role in defining brand similarities.
• Further analysis with larger datasets could yield more robust insights into cigarette brand characteristics and their public health implications.
• Integrated data analysis reveals complex interdependencies between cigarette constituents, market factors, and public health indicators.