第19期青年學(xué)者金融論壇
題目：Explore the link between fine grained energy data and household characteristics for small area census
報(bào)告人：Sharon Xiaowen Lin博士（南安普頓大學(xué)研究員）
主持人：馮小兵教授
時(shí)間：2015年4月14日下午13：00
地點(diǎn)：博萃樓陽光房
論文：學(xué)院內(nèi)網(wǎng)
Sharon Xiaowen Lin簡介：
Bio of Dr. Lin
Dr Sharon Xiaowen Lin is a research fellow University of Southampton, working on the frontier of the interdisciplinary research in the area energy and climate change. She is currently a member of Sustainable Energy Research Group and Southampton Statistical Science Research Institute. She has been awarded several fellowships and research projects from UK research funding bodies and government. She was a guest research fellow at Parliamentary Office of Science and Technology (2009-2010). Dr Lin studied MSc economics at University of London and PhD at Cass Business School, London. She has taught and supervised students at various levels including PhD at Cass Business School and served as a referee for a number of peer-reviewed journals. Her published papers have appeared in various SCI Indexed journals such as Energy Economics, Energy Policy.
論文摘要：
Explore the link between fine grained energy data and household characteristics for small area census
B. Ansderson
University of Southampton, Southampton, UK (B.Anderson@soton.ac.uk)
Sharon Xiaowen Lin
University of Southampton, Southampton, UK (X.Lin@soton.ac.uk)
Andy Newing
University of Leeds, Leeds, UK (a.newing@leeds.ac.uk)
2021 Census is likely to be the last UK decennial census since 1801. There is the need for robust alternative methods to produce small area social-economic indicators over time. This research explores the possibility of using fine grained energy consumption data (half-hourly data) to come up with small area ‘census-like’ indicators. We use multilevel modelling techniques to over 4000 Irish households to establish the relationship between energy consumption and household characteristics, followed by regression methods to examine census-like indicators (e.g. income, employment, floor area) using energy consumption data at the household level. Our preliminary results indicate there is a link between energy consumption and household characters and there are challenges for fine grained energy data (extremely large data, mismatched frequencies) to be used as a sensible candidate for census-like indicators.
Keywords: fine grained energy data, big data, census-like indicators, mixed-effect modelling
