White House Energy Datapalooza: PlotWatt

PlotWatt helps people reduce their energy bills by providing customized money-saving recommendations.

Transcript:

1
00:00:01,166 --> 00:00:03,099
Speaker:
All right.

2
00:00:03,100 --> 00:00:05,567
Meet Susan.

3
00:00:05,567 --> 00:00:08,700
Susan recently bought a house.

4
00:00:08,700 --> 00:00:11,265
Nice house, Susan.

5
00:00:11,266 --> 00:00:16,166
Unfortunately, Susan's house
came with electric bills,

6
00:00:16,166 --> 00:00:20,900
bills that kept
growing out of control.

7
00:00:20,900 --> 00:00:23,299
So she decided to
do some research.

8
00:00:25,467 --> 00:00:30,967
But everywhere she looked, she
found frustration and confusion.

9
00:00:30,967 --> 00:00:33,967
Walking down the aisles
of a hardware store,

10
00:00:33,967 --> 00:00:38,967
Susan saw adds that said she
could save up to $500 with a new

11
00:00:38,967 --> 00:00:44,333
water heater, but she was pretty
sure her water heater was new,

12
00:00:44,333 --> 00:00:47,266
but maybe she had the
wrong new water heater.

13
00:00:47,266 --> 00:00:51,934
You see, Susan didn't know
that her existing appliances,

14
00:00:51,934 --> 00:00:55,266
whether she bought them
yesterday or 30 years ago,

15
00:00:55,266 --> 00:00:58,766
were much smarter
than she thought.

16
00:00:58,767 --> 00:01:05,133
You see, all of our appliances
are chattering away constantly.

17
00:01:05,132 --> 00:01:08,265
They're chattering
away at the same time.

18
00:01:10,734 --> 00:01:15,767
The problem is that our
appliance translator is broken.

19
00:01:15,767 --> 00:01:20,000
For more than a century, we have
looked at our electricity meters

20
00:01:20,000 --> 00:01:23,734
as cash registers for
utility companies,

21
00:01:23,734 --> 00:01:28,399
when they are capable
of being so much more.

22
00:01:28,400 --> 00:01:33,400
At Plotwatt, we develop software
and we write algorithms.

23
00:01:33,400 --> 00:01:37,867
Algorithms that take this data
from your existing smart meter

24
00:01:37,867 --> 00:01:41,533
and translate what your
appliances are doing.

25
00:01:41,533 --> 00:01:46,900
You see, each one of your
appliances has a fingerprint,

26
00:01:46,900 --> 00:01:50,066
a fingerprint that's
characterized by things like

27
00:01:50,066 --> 00:01:53,934
power consumption over time,
but there's much more to an

28
00:01:53,934 --> 00:01:56,467
appliance fingerprint.

29
00:01:56,467 --> 00:01:59,767
We've been working since 2008
with our team of scientists and

30
00:01:59,767 --> 00:02:02,800
mathematicians and software
developers to figure out what

31
00:02:02,800 --> 00:02:06,265
goes into an
appliance fingerprint.

32
00:02:06,266 --> 00:02:10,667
And now what we have is we have
the ability to disaggregate

33
00:02:10,667 --> 00:02:11,667
energy consumption.

34
00:02:11,667 --> 00:02:15,166
We are actively doing this on
thousands of homes and chain

35
00:02:15,166 --> 00:02:17,867
restaurants around the country.

36
00:02:17,867 --> 00:02:19,033
This is cool.

37
00:02:19,033 --> 00:02:23,065
This is an itemized
electric bill.

38
00:02:23,066 --> 00:02:27,333
What's really cool, though, is
what you can do with the data

39
00:02:27,333 --> 00:02:31,266
once you know what's happening
every minute of the day

40
00:02:31,266 --> 00:02:32,934
in a home.

41
00:02:32,934 --> 00:02:36,266
It enables you to give
people customized,

42
00:02:36,266 --> 00:02:39,333
quantified recommendations.

43
00:02:39,333 --> 00:02:43,934
What if you could know when it's
actually time to clean the coils

44
00:02:43,934 --> 00:02:48,066
on your refrigerator, which I'm
sure none of us has done any

45
00:02:48,066 --> 00:02:52,000
time recently, but if we could
quantify exactly how much that

46
00:02:52,000 --> 00:02:55,700
would save you, that
would be meaningful.

47
00:02:55,700 --> 00:02:58,700
So back to Susan.

48
00:02:58,700 --> 00:03:03,399
Susan read about Plotwatt
in some Datapalooza press,

49
00:03:03,400 --> 00:03:07,367
went to her utility website,
connected her data with

50
00:03:07,367 --> 00:03:13,333
Plotwatt, and we ran our crude
anomaly detection algorithms

51
00:03:13,333 --> 00:03:17,367
over the data, and
immediately revealed insights.

52
00:03:17,367 --> 00:03:19,834
You see, Susan suffered from
a problem that lots of our

53
00:03:19,834 --> 00:03:23,667
existing users suffer from,
her thermostat was telling her

54
00:03:23,667 --> 00:03:27,233
auxiliary heat to come on
when it shouldn't have been.

55
00:03:27,233 --> 00:03:30,800
This is something that we see
happening time and time again,

56
00:03:30,800 --> 00:03:35,967
it's one little button that
can save Susan big bucks.

57
00:03:35,967 --> 00:03:40,400
In fact, this is hundreds of
dollars in Susan's pocket in the

58
00:03:40,400 --> 00:03:42,367
winter months.

59
00:03:42,367 --> 00:03:45,767
So this is what we built
for the Datapalooza,

60
00:03:45,767 --> 00:03:47,734
this is a section of the site.

61
00:03:47,734 --> 00:03:51,967
We're happy to show you guys the
full green button experience in

62
00:03:51,967 --> 00:03:55,033
the demo session, but this is
what we call the check engine

63
00:03:55,033 --> 00:03:57,166
light for your house.

64
00:03:57,166 --> 00:04:00,233
You upload your
green button data or,

65
00:04:00,233 --> 00:04:02,734
hopefully in the future
connect your data,

66
00:04:02,734 --> 00:04:06,300
and we run a crude anomaly
detection algorithm over the

67
00:04:06,300 --> 00:04:09,800
data to look for any
strange performance.

68
00:04:09,800 --> 00:04:13,433
In this case, we pop up
a warning, a red light,

69
00:04:13,433 --> 00:04:16,466
a check engine light for her
home that says something is

70
00:04:16,466 --> 00:04:18,899
wrong with your heating and
air conditioning system,

71
00:04:18,899 --> 00:04:20,767
you should check it out.

72
00:04:20,767 --> 00:04:24,033
This is cool, this is what we
can do with 15-minute data or

73
00:04:24,033 --> 00:04:25,667
one-hour data.

74
00:04:25,667 --> 00:04:29,400
What's really cool, though, is
what we can do when we unleash

75
00:04:29,400 --> 00:04:32,332
the full potential
of our smart meters.

76
00:04:32,333 --> 00:04:35,934
If we can get a data
sample a minute,

77
00:04:35,934 --> 00:04:41,400
not only do we multiply the
appliances that we can pull out,

78
00:04:41,400 --> 00:04:46,166
but more importantly, we
multiply the types of insight

79
00:04:46,166 --> 00:04:48,033
that we can provide.

80
00:04:48,033 --> 00:04:54,834
For instance, most of us either
forget to or actually change the

81
00:04:54,834 --> 00:04:57,734
filters in our HVAC
system on a schedule.

82
00:04:57,734 --> 00:05:00,667
It's a rule of thumb.

83
00:05:00,667 --> 00:05:04,066
There's no reason why we should
be changing our HVAC filters on

84
00:05:04,066 --> 00:05:05,265
a schedule.

85
00:05:05,266 --> 00:05:08,133
If we can get a sample a minute,
we can tell you when you should

86
00:05:08,133 --> 00:05:11,700
actually be changing the
filters on your HVAC.

87
00:05:11,700 --> 00:05:14,233
Maybe there is systemic
opportunity in your house that's

88
00:05:14,233 --> 00:05:17,467
much larger than that, but this
is the way that we should be

89
00:05:17,467 --> 00:05:18,567
thinking about our homes.

90
00:05:18,567 --> 00:05:21,667
Our homes are systems
that can be optimized,

91
00:05:21,667 --> 00:05:24,833
and when we just unleash the
full potential of our existing

92
00:05:24,834 --> 00:05:27,066
smart meters, this is
no additional hardware,

93
00:05:27,066 --> 00:05:29,265
this is just the data
from smart meters,

94
00:05:29,266 --> 00:05:33,467
this is the sort of
thing that we can do.

95
00:05:33,467 --> 00:05:37,834
No more rules of thumb,
no more guesstimations,

96
00:05:37,834 --> 00:05:40,367
no more marketing mumbo-jumbo.

97
00:05:40,367 --> 00:05:45,233
Plotwatt gives you actionable,
customized, quantified insight.

98
00:05:45,233 --> 00:05:49,033
This is the way
energy data should be.

99
00:05:49,033 --> 00:05:52,266
Thank you for having us, and a
special thanks to Pecan Street

100
00:05:52,266 --> 00:05:54,834
for helping out with
our green button app.

101
00:05:54,834 --> 00:05:57,567
(applause)