Small Area Estimation as a tool for thinking about spatial variation in energy demand

Small Area Estimation as a tool for thinking about spatial variation in energy demand Dr Ben Anderson Sustainable Energy Research Centre University of Southampton @dataknut Spatial Variation in Energy Use, Attitudes and Behaviours: Implications for Smart Grids and Energy Demand Policy Studies Institute, London: 7 February 2014

Small Area Estimates of Electricity Consumption Contents  What & Why  How?  Results – Overall consumption – Consumption inequalities  Conclusions & future Directions @dataknut 2

Small Area Estimates of Electricity Consumption Contents  What & Why  How? ?  Results – Overall consumption – Consumption inequalities  Conclusions & future Directions @dataknut 3

Small Area Estimates of Electricity Consumption Digression: Geography  Southampton (UK) @dataknut 4

Small Area Estimates of Electricity Consumption Digression: What‟s a small area?  In this case… – English Lower Layer Super Output Areas – Census 200/2011 LSOAs – c. 630 households each – 148 in Southampton City @dataknut 5

Small Area Estimates of Electricity Consumption What & Why  Basically we want something for nothing – Small area estimates of energy demand – Without a bespoke energy census  Why? – – – – @dataknut Infrastructure planning Energy efficiency intervention analysis Energy inequality analysis Politics! 6

Small Area Estimates of Electricity Consumption The problem:  Small area summaries exist  But they are aggregates – Or averages – Or profiled  And we want a micro-level simulation model to assess the socio-economic and spatial impact of Price changes Incentive changes ‘Efficiency’ interventions Changes in ‘energy habits’ (appliances & practices) Socio-demographic change @dataknut 7

Small Area Estimates of Electricity Consumption What can we do?  A bespoke energy census – ££££££££££££££££££££ @dataknut 8

Small Area Estimates of Electricity Consumption What can we do?  A bespoke energy census – ££££££££££££££££££££  A large sample energy survey covering all LSOAs – ££££££££££ @dataknut 9

Small Area Estimates of Electricity Consumption What can we do?  A bespoke energy census – ££££££££££££££££££££  A large sample energy survey covering all LSOAs – ££££££££££  Small Area Estimation – Take existing area level data – Take (ideally) an existing large n survey – Combine £ @dataknut 10

Small Area Estimates of Electricity Consumption Small Area Estimation  Econometric approaches Income, income deprivation, income inequality smoking prevalence, obesity, consumption expenditure, CO2, water… – Well known – Multi-level Models – Usually requires census microdata for anything other Innovation Network: than means “Evaluating and improving small area estimation methods”  Re-weighting (and other) approaches – Increasingly http://eprints.ncrm.ac.uk/3210/ well known – 'Spatial microsimulation' – Does not require census microdata @dataknut 11

Small Area Estimates of Electricity Consumption Contents  What & Why Estimation  How?  Results – Overall consumption – Consumption inequalities  Conclusions & Future Directions @dataknut 12

Small Area Estimates of Electricity Consumption Data  Data – Living Costs and Food Survey 2008-2010 Consumption proxies (reported energy expenditure) – Census 2001 (2011)  Projection/forecasting – Projected ‘surveys’ -> 2021 – Projected ‘census’ -> 2021  So far… @dataknut 13

Small Area Estimates of Electricity Consumption Conceptually… LSOA census ‘constraint’ tables Survey data cases If Region = 1 LSOA 1.1 (Region1) If Region = 2 @dataknut Weights LSOA 2.1 (Region2) Iterative Proportional Fitting Ballas et al (2005) 14

Small Area Estimates of Electricity Consumption Key First Job:  Choose your constraints Census data Survey data  You may have little choice @dataknut 15

Small Area Estimates of Electricity Consumption Key First Job: LSOA census ‘constraint’ tables  The constraints – Selected by stepwise regression Expenditure Number of persons Employment Status Accommodation type Number of earners Age of HRP Age of HRP Employment Status Most important Share of expenditure Composition Number of rooms Number of children Least important R sq @dataknut Ethnicity (non-white) 0.136 0.01 16

Small Area Estimates of Electricity Consumption „Iterative Proportional Fitting‟  Well known!  Deming and Stephan 1940 – Fienberg 1970; Wong 1992  A way of iteratively adjusting statistical tables – To give known margins (row/column totals) – ‘Raking’  In this case – Create weights for each case so LSOA totals ‘fit’ constraints – Weighting ‘down’ @dataknut 17

Small Area Estimates of Electricity Consumption Internal Validation methods  Use of constraints to re-create the Census tables  Difference = Absolute Error – – Total Absolute Error (TAE) = sum of all errors Standardised AE = TAE/(n persons x n constraint categories)  Smith et al: – SAE of less than 20% and ideally less than 10% – Consumption Mean SAE p90 Ethnicity 2.18% 3.05% Number of children 0.11% 0.22% Number of rooms 0.05% 0.10% Employment status (HRP) 0.88% 1.22% Age (HRP) 0.34% 0.75% Tenure 0.07% 0.14% Accomodation type 0.21% 0.51% Number of persons 0.00% 0.00% in 90% of the areas is desirable. @dataknut 18

Small Area Estimates of Electricity Consumption Preliminary results: Electricity  Mean weekly household £  Modelled  Census 2001  LC&F Survey 20082010 @dataknut 19

Small Area Estimates of Electricity Consumption Validation: Electricity  Mean weekly household £  Observed @LSOA – DECC 2010  Spearman: 0.317 @dataknut 20

Small Area Estimates of Electricity Consumption Preliminary results: Electricity  Total weekly household £  Modelled  Census 2001  LC&F Survey 2008-2010 @dataknut 21

Small Area Estimates of Electricity Consumption Validation: Electricity  Total weekly household £  Observed @LSOA – DECC 2010  Spearman: 0.509 @dataknut 22

Small Area Estimates of Electricity Consumption What is causing the error?  Heating! – @dataknut 2011 data 23

Small Area Estimates of Electricity Consumption What is causing the error?  Housing growth  Combined: – – @dataknut Heating = 60% Growth = 5% 24

Small Area Estimates of Electricity Consumption Consumption inequality  Area level gini  £ mean spend  R = -0.413 – @dataknut (p < 0.001) 25

Small Area Estimates of Electricity Consumption Consumption inequality  Area level gini  Index of Multiple Deprivation 2010 – Income score  R = 0.463 – @dataknut (p < 0.001) 26

Small Area Estimates of Electricity Consumption My big worry  Data quality Source: SPRG/ARCC-Water Survey, 2011 www.sprg.ac.uk @dataknut 27

Small Area Estimates of Electricity Consumption Contents  What & Why  How?  Results – Overall consumption – Consumption inequalities  Conclusions & Future Directions @dataknut 28

Small Area Estimates of Electricity Consumption Conclusions  Outliers and errors are informative  Reported consumption data – Could be dangerous  Census 2011 central heating – Critical new constraint @dataknut 29

Small Area Estimates of Electricity Consumption Future directions  Update for 2011 data  Census projection 1981 -> 2021  Use measured energy consumption – New survey data?  Contact: – b.anderson@soton.ac.uk @dataknut 30

Small Area Estimation as a tool for thinking about spatial variation in energy demand

by Ben Anderson, Senior Research Fellow, Sustainable Energy Research Group at University of Southampton on Feb 12, 2014

Statistics

Views

Actions

1 Embed 1

Accessibility

Categories

Upload Details

Usage Rights

Report content

Small Area Estimation as a tool for thinking about spatial variation in energy demand Presentation Transcript