27 / 28
Fuel Poverty: Socioeconomic and political aspects
J. PARAVANTIS
175
the country’s threshold, with a high fraction of households not using heating energy
at all. Finally, Paravantis and Santamouris
34
used k-means clustering to group the 50
households into three clusters (Poorest, Average and Richest) based on mean indoor
temperature, surface area of the dwelling, number of rooms, family size, building age
and income. They found that 7.6% of the households of the Richest Cluster, 8.6% of the
Average Cluster and 11.6% of the Poorest Cluster were fuel poor with indoor tempera-
tures registering much below accepted standards.
III. Methodology
The following research questions have been answered by the aforementioned three
journal papers for the case of Greece (a rather typical Mediterranean country):
1. Which social, economic and physical/infrastructure variables are influential in
grouping low-income households into homogeneous clusters? How many such clusters
are formed? Do these clusters correspond to social/income classes? How do the clus-
ters compare with one another in terms of available living space, employment status,
income, household insulation, energy consumption and fuel poverty?
2. How do the measured indoor temperatures vary per month, day of the week, hour,
household and cluster? How may the temperature time series be modeled as a function
of time, season and the socioeconomic characteristics of the household?
Two levels of analysis were employed: 598 low income households were analyzed via
questionnaires for the successive winters of 2010-2011 (among the warmest winters
on record in Greece, dating back to the 19th century) and 2001-2012 (among the 15%
of coldest winters on record ever); and indoor temperatures were measured in 50 low-
and very-low income dwellings in Athens, Greece during the winter of 2012-2013.
IV. Results
On the results of the analysis of the 598 households,
35
a survey held in the spring and
summer of 2012 collected data of the heating energy consumption for 2010-2011 and
2011-2012, from 598 households via a 2-page, 50-point questionnaire. All but the low-
est income classes were found to have lost 12.7 to 31% of their 2009 income while
the lowest income class gained about a fourth of its 2009 income, the distribution
of which is depicted in Figure 1, likely because more household members joined the
34. Paravantis, J. A. and Santamouris, M., An analysis of indoor temperature measurements in low and very
low income housing in Athens, Greece. Special Issue on Indoor Environmental Quality in Low Income
Housing in Europe of the Journal
Advances in Building Energy Research
, Taylor and Francis, accepted
for publication, 2015.
35. Santamouris, M., Paravantis, J. A., Founda, D., Kolokotsa, D., Michalakakou, P., Papadopoulos, A. M.,
Kontoulis, N., Tzavali, A., Stigka, E. K., Ioannidis, Z., Mehilli, A., Matthiessen, A. and Servou, £., Financial
crisis and energy consumption: A household survey in Greece. Energy and Buildings, 2013, 65, 477-487.