@inproceedings{QuinnTerziCrovella:KDD22,
  author	= {Quinn, Kevin and Terzi, Evimaria and Crovella, Mark},
  title		= {Characterizing Covid Waves via Spatio-Temporal Decomposition},
  booktitle	= {Proceedings of KDD},
  URL		= {http://www.cs.bu.edu/faculty/crovella/paper-archive/kdd22-covid-waves.pdf},
  doi		= {10.1145/3534678.3539136},
  address	= {Washington, DC},
  year		= {2022},
  abstract	= {In this paper we develop a framework for analyzing prevalence data of a disease such as Covid. Given such data we are able to identify the waves (temporal patterns) that are present in the data and the spatial epicenters of each wave. In order to achieve that, we introduce a new spatio-temporal decomposition of the data which we call diffusion NMF (D-NMF). The key characteristic of D-NMF is that it takes into consideration the spatial structure of the waves such that locations that are spatially close are more likely to experience the same wave. We use D-NMF to extensively analyze Covid prevalence data at various spatial granularities. Our results demonstrate that D-NMF is very useful in separating the waves of the epidemic and identifying a few centers for each wave.}
}