Slide 1
Slide 1 text
Decoding the SAVE project data
Tom Rushby
[email protected]
@tom_rushby
Ben Anderson
[email protected]
@dataknut
Slide 2
Slide 2 text
Menu
§ Data collection
– Loop data journey
– What else are we collecting?
– Constructing the dataset
2
§ Exploring the data
– Become an energy analyst
§ What next?
Image (modified): ‘matrix’ by
Gamaliel Espinoza Macedo
https://www.flickr.com/photos/gamikun/
Slide 3
Slide 3 text
§ Data collection
– Loop data journey
– What else are we collecting?
– Constructing the dataset
Menu
3
Image (modified): ‘matrix’ by
Gamaliel Espinoza Macedo
https://www.flickr.com/photos/gamikun/
§ Exploring the data
– Become an energy analyst
§ What next?
Slide 4
Slide 4 text
Data collection
4
§ Loop data journey (1)
Photos by Patrick James | [email protected] | @pab_james
Slide 5
Slide 5 text
Data collection
5
§ Loop data journey (2)
Slide 6
Slide 6 text
Data collection
6
§ Loop data journey (3)
Slide 7
Slide 7 text
Data collection
7
Energy Loop Recruitment survey
Time-use diaries Update survey
SAVE
ý
þ
þ
þ
§ What are we collecting? Presenting today?
Slide 8
Slide 8 text
Constructing the dataset
8
Analysis and reports
Survey
updates
Recruitment
survey
Energy Loop
data
Slide 9
Slide 9 text
Constructing the dataset
9
§ How much data?
Ø Meter Clamps
o (4,318 households)
o 15 minute consumption (Wh)
à 414,000 records/day
o 10 second power (W)
à 37m records/day
Ø Collecting for ~3 years
o ~2 Terabytes
Slide 10
Slide 10 text
Menu
10
§ Data collection
– Loop data journey
– What else are we collecting?
– Constructing the dataset
§ Exploring the data
– Become an energy analyst
§ What next?
Slide 11
Slide 11 text
Exploring the data
§ We will use different time scales to reveal
patterns in consumption
Today we will look at data collected over:
• Months
• Weeks
• Hours
11
Slide 12
Slide 12 text
Exploring the data
12
0.06
0.09
0.12
0.15
Thu May 19
Thu Jun 02
Thu Jun 16
Thu Jun 30
Thu Jul 14
Thu Jul 28
Thu Aug 11
Thu Aug 25
Thu Sep 08
Thu Sep 22
Thu Oct 06
Thu Oct 20
Thu Nov 03
Thu Nov 17
Thu Dec 01
Thu Dec 15
Thu Dec 29
Thu Jan 12
Thu Jan 26
Thu Feb 09
Thu Feb 23
Thu Mar 09
kWh
Let’s look at this period
0.06
0.09
0.12
0.15
Thu May 19
Thu Jun 02
Thu Jun 16
Thu Jun 30
Thu Jul 14
Thu Jul 28
Thu Aug 11
Thu Aug 25
Thu Sep 08
Thu Sep 22
Thu Oct 06
Thu Oct 20
Thu Nov 03
Thu Nov 17
Thu Dec 01
Thu Dec 15
Thu Dec 29
Thu Jan 12
Thu Jan 26
Thu Feb 09
Thu Feb 23
Thu Mar 09
Date
kWh
Mean Median
Data: Navetas Loop 15m (2016−06−11 to 2017−03−05, all SAVE data received)
Mean of 15 minute intervals
Slide 13
Slide 13 text
13
Exploring the data
Let’s look at this period
Slide 14
Slide 14 text
Exploring the data
14
Slide 15
Slide 15 text
15
Exploring the data
Slide 16
Slide 16 text
16
Exploring the data
Slide 17
Slide 17 text
17
Exploring the data
Slide 18
Slide 18 text
18
Exploring the data
Slide 19
Slide 19 text
Menu
19
§ Data collection
– Loop data journey
– What else are we collecting?
– Constructing the dataset
§ Exploring the data
– Become an energy analyst
§ What next?
Slide 20
Slide 20 text
What next?
More analysis (and data) to understand:
§ how households respond to different
incentives across time
§ the characteristics of those who
respond in the same way
§ how energy-using habits change
20
Slide 21
Slide 21 text
What next?
More data!
Using time-use data
to better understand
changes in energy
using practices or
routines
21
% point change in timing of reported laundry by weekday vs
weekend and by time of day for each year
(MTUS UK sub-sample weighted, for clarity, only 95% confidence
intervals for Sunday included).[1]
1 Anderson, Ben. ‘Laundry, Energy and Time: Insights from 20 Years of Time-Use Diary
Data in the United Kingdom’. Energy Research & Social Science 22 (2016): 125–36.
doi:10.1016/j.erss.2016.09.004.
Slide 22
Slide 22 text
Thank you for listening.
22
Photo: ‘Learning’ by CollegeDegrees360
used under Creative Commons License