Profile picture
Casey Wichman @cjwich
, 16 tweets, 5 min read Read on Twitter
My wife and I have a monthly debate about whether our @nest thermostat uses more/less energy. My argument has always been: probably a null effect, but it's fun technology. Her argument: we're spending more money on this toy and our bills seem higher. [1/n]
Her argument is fair, but I've contended that our monthly Home Report from @nest simply makes our electricity consumption more salient (see, e.g., rff.org/research/publi…) [2/n]
So, I pulled our historical hourly data from @pepco and plotted the results: [3/n]
I find it hilarious that we peak during the winter between xmas and the @ASSAMeeting, typically when we're working from home and winter has just set in. [4/n]
Context: we live in a small 1BR apartment in DC, electric heat, central air. Prior to our smart thermostat, we've had a dumb one. [5/n]
After controlling for max daily temp and humidity, presence of rain, and absorbing month and day-of-week FEs, our residual electricity consumption looks pretty flat before and after our thermostat replacement. [6/n]
Add in a dummy for whether we had our smart thermostat installed, we can look at its impact on our daily electricity (in kwh). Results suggest that the @nest thermostat increased our energy consumption by ~1.5 kwh/day, but the effect is statistically zero. [7/n]
if you're worried that electricity consumption is distribution log-normally, I got you... magnitude is large (~5% increase in consumption), but not significantly different from zero [8/n]
...and the effect holds up if we use only the months for which we have pre-smart thermostat data. [9/n]
For those p-value hawks out there, recall this is an n=1 and I'd like to think that we're a fairly select sample 😎. [10/n]
Working title: "Do smart thermostats save energy? Evidence from artisanal data" [11/n]
It's easy to joke, but these thermostats sell for $250 a pop claiming energy conservation and I've seen ZERO pieces of causal evidence that they actually save energy, and there are good arguments for why usage might go up. [12/n]
I like my thermostat, but not because I think it saves me energy. Sorry to be harsh, @Nest, but if you want to claim energy savings, prove it. I know an economist or two who'd be happy to partner with you to run some experiments 😉 [n/n]
To all y'all suggesting alternative weather specifications, I think you're missing the point: the @nest was an xmas gift so it's orthogonal to weather. But you're right, I'll implement an ML #algo to sort through weather specifications until p<0.01

[(n+1)/n]
At this point, it’s worth re-emphasizing the point of econometrics: we use sample data to infer what’s going on in a more general population. My electricity use is not representative and should not be used as evidence of whether smart thermostats work in general [(n+2)/2]
But it’s also worth emphasizing that I have not seen large-scale, *causal* evidence in one direction or the other. So, I’ll keep wearing my economist hat of “zero until proven otherwise.” [(n+3)/n]
Missing some Tweet in this thread?
You can try to force a refresh.

Like this thread? Get email updates or save it to PDF!

Subscribe to Casey Wichman
Profile picture

Get real-time email alerts when new unrolls are available from this author!

This content may be removed anytime!

Twitter may remove this content at anytime, convert it as a PDF, save and print for later use!

Try unrolling a thread yourself!

how to unroll video

1) Follow Thread Reader App on Twitter so you can easily mention us!

2) Go to a Twitter thread (series of Tweets by the same owner) and mention us with a keyword "unroll" @threadreaderapp unroll

You can practice here first or read more on our help page!

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just three indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member and get exclusive features!

Premium member ($3.00/month or $30.00/year)

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!