Speaker
Description
The purpose of this paper is to present a novel methodology for profiling household sustainable consumption scores. The profiling starts with initial clusterisation with 3 clusters. The analysis continues with profiling of clusters by the main segmentations criteria (sub-segmentation). The used dataset contains 2117 rows. These are households from Varna, Sofia and Svishtov, surveyed in 2021 and 2022. The scores are calculated in 5 dimensions (total score, housing, food, mobility and product disposal) for each household. The multiple regression model cannot be applied for profiling because of high levels of correlation among the dimensions of sustainable consumption (multi-collinearity). Moreover, the functional dependency between the total score and its dimensions decreases the number of methods which may be used for profiling. The profiling procedures are closely related to clustering. The well-known techniques such as k-means clustering may reveal some clusters of sustainable consumption activities. But having 4 dimensions of sustainable consumption (housing, food, mobility and product disposal) outlines different profiles. Starting from the total score, the categorization may be done by 3 categories: high, middle and low levels of the total score. These 3 categories may be applied to each category: housing, food, mobility and product disposal. When categorizing the dataset by each category, people with a high total score may have high scores in all four dimensions; however, it may be a situation where respondents with a high total score have lower scores in the four dimensions. This approach outlines nuances in the people with high, middle and low total scores. What kind of combinations exist in the three dimensions of each level of total sustainable consumption? The proposed approach will give the possibility for detailed (in-depth) analysis for the sources to reach the expected results of the total sustainability index. Visualisation of the results is given. It is useful for the visualisation of similar approaches for clustering. Similar analysis may be conducted with other datasets for sustainable consumption. Other researchers may adapt the proposed methodology. Further steps of the data analysis following the interdisciplinary spiral include the identification of common characteristics in the identified complex categories. This research is financed by the project NPI-70/2023 “Sustainable consumer activities of Varna’s households – possible reasons”.
