Multi-Method Approach to Compare the Socio-Demographic Typology of Residents and Clusters of Electricity Load Curves in a Swiss Sustainable Neighbourhood
Keywords:
applied statistics, typology, energy, cluster, sustainable consumption, researchAbstract
A sustainable neighbourhood was built Switzerland by one of the leaders in this field. Half of the 400 apartments have been equipped with smart meters delivering big data on energy consumption (electricity, water, heating…). The company would like to know if it is possible to link socio-demographic typology of residents with energy consumption patterns. To answer this question we present in this article a multi-method approach combining qualitative analysis, frequently used in marketing (multiple correspondence analyses), and quantitative analysis from applied statistics to answer this question. First, we have conducted a survey among the residents of the sustainable neighbourhood to gather socio-demographic data, and then we have proposed a marketing typology of residents. In parallel, we have analysed load curves with statistical models (clustering factors, hermano beta models, coincidence factors, som, expert practice) to see if there are patterns of energy consumption and to determine groups of similar load curves. Then we have compared the discrepancies in the composition of the groups between both methods. This study is based on a single case study generating a new research hypothesis: the typology of residents based on socio-demographic data can be linked to energy consumption pattern of a household.
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
References
Abdi, H., Valentin, D. (2007), "Multiple correspondence analysis", Encyclopaedia of measurement and statistics, pp 651–657.
Dudek, A., Hurd, H., Wojtowicz, W., Wojtowicz, M.W. (2013), "Package ‘per ARMA",http://watson.nci.nih.gov/cran_mirror/web/packages/perARMA/perARMA.pdf (13/5/2016)
Josse, J., Chavent, M., Liquet, B., Husson, F. (2012), "Handling missing values with regularised iterative multiple correspondence analysis", Journal of classification. 29 (1), 91–116.
Montero, P., Vilar, J.A. (2014), "TSclust: An R Package for Time Series Clustering", http://www.jstatsoft.org/v62/i01/paper (22/9/2015)
Wehrens, R., Buydens, L.M. et al. (2007), "Self-and super-organising maps in R: the Kohonen package". J Stat Softw., 21 (5), 1–19.
Gower, J., Lubbe, S. Roux, N. l. (2011), “Multiple Correspondence Analysis”, in Understanding Biplots, John Wiley & Sons, Ltd, Chichester, UK.doi: 10.1002/9780470973196.ch8
Yates, F. (1934), “Contingency Tables Involving Small Numbers and the χ2 Test” Supplement to the Journal of the Royal Statistical Society, 1(2), 217–235, available at: http://doi.org/10.2307/2983604 (5/5/2016)
