STRAIGHTFORWARD APPROACH IN CULTURAL HERITAGE AND ENVIRONMENT STUDIES - MULTIVARIATE ANALYSIS AND CHEMOMETRY Cluster analysis of microclimate data to optimize the number of sensors for the assessment of indoor environment within museums Anna Maria Siani 1 & Francesca Frasca 2 & Marta Di Michele 1 & Valerio Bonacquisti 3 & Eugenio Fazio 3 Received: 23 November 2017 /Accepted: 11 April 2018 # Springer-Verlag GmbH Germany, part of Springer Nature 2018 Abstract For the first time, the cluster analysis (k-means) has been applied on long time series of temperature and relative humidity measurements to identify the thermo-hygrometric features in a museum. Based on ASHRAE (2011) classification, 84% of time all rooms in the Napoleonic Museum in Rome (case study) were found in the class of control B. This result was obtained by analyzing all recorded data in 10 rooms of the museum as well as using the cluster aggregation. The use of objective-oriented methodology allows to achieve an acceptable knowledge of the microclimate in case of multi-room buildings, reducing com- putations with large amounts of collected data and time-consuming in redundant elaborations. The cluster analysis enables to reduce the number of the sensors in microclimate monitoring programs within museums, provided that the representativeness of the instrument location is known, and professional conservators have assessed that the artifacts are well preserved. Keywords Data quality assessment . Cluster analysis . Temperature . Relative humidity . Museum Introduction The conservation of artworks and collections is strongly af- fected by environmental conditions, which may be directly and indirectly responsible for their deterioration processes. In the last 30 years, a growing number of studies has been conducted with the aim at defining the adequate microclimate to better preserve the cultural heritage. The first works (Bernardi and Camuffo 1995; Camuffo and Bernardi 1996; Pavlogeorgatos 2003) were focused on char- acterizing the indoor climate in terms of temperature and rel- ative humidity averages and their variability, as well as on investigating the impact of external climate conditions on indoors (Camuffo et al. 1999). The recent methodological approach is instead oriented to determine the historic climate (i.e., the climate to which the object has been acclimatized and satisfactorily been preserved for a long time). In the case of objects sensitive to climate-induced damage, the approach is to apply the European Norm EN 15757 (2010), which recom- mends establishing the safety range of relative humidity if this climate has been proved to be beneficial for the conservation of preserved objects by qualified specialists (Camuffo 2014; Bertolin et al. 2015). In the last decade, more attention has been also paid to understand the causes of damage in the artwork triggered by climate conditions (Bratasz et al. 2007). In case of multi-room buildings, the assessment of indoor climate is performed analyzing thermo-hygrometric data re- corded in one or more rooms in which the most valuable artworks are preserved. This is for two main reasons: (1) it is not possible to install instruments in each room due to cost sustainability and (2) indoor climates of contiguous rooms may be similar, and consequently, several sensors might be useless. However, when several sensors are placed in the same room or in every room on the same floor, a reduction of the same could be strategic in terms of maintenance and calibra- tion costs. Consequently, it is possible to optimize the site configuration setting using a minor number of sensors whose Responsible editor: Constantini Samara * Anna Maria Siani annamaria.siani@uniroma1.it 1 Department of Physics, Sapienza Università di Roma, P.le A. Moro 2, 00185 Rome, Italy 2 Department of Earth Sciences, Sapienza Università di Roma, P.le A. Moro 2, 00185 Rome, Italy 3 Department of Fundamental and Applied Science for Engineering, Sapienza Università di Roma, P.le A. Moro 2, 00185 Rome, Italy Environmental Science and Pollution Research https://doi.org/10.1007/s11356-018-2021-3