Research Fellow - Data Scientist, Centre for Social Data Analytics (CSDA), School of Economics
Phone: +64 9 921 9999 – ext: 5097
Chamari has over 10 years’ experience of working on data science, computer science, operational research, and statistics related research projects in universities, including several years as a full-time lecturer.
At CSDA, Chamari is working on predictive risk modelling research projects which aim to predict the risks associated with homelessness or likelihood of being a victim of a child maltreatment event. She is the lead Data Scientist of Allegheny County’s Homeless and Mental Health Housing Service Prioritization research projects. Chamari has contributed to Douglas County child welfare decision aid by assuring the quality of data to be modelled, performing external validation, and exploring alternative state-of-art ensemble decision models.
Furthermore, Chamari is interested in enhancing the throughput of high speed, concept drifting data stream classifiers without sacrificing its accuracy. In her PhD, she proposes a novel stage learning framework that senses the context of data to determine the level of volatility in the stream. In addition, Chamari introduces an innovative decision tree forest driven incremental Fourier classifier ensemble for classification in evolving data streams.
In her previous research engagements, she worked with facility location optimisation problems, and an e-Business collaboration modelling. During her tenure as a computer Science lecturer in Sri Lanka and New Zeeland, Chamari has worked as the primary supervisor for several undergraduate projects.