In a study published 20th February 2020 by tado they found that UK homes loss heat considerable faster than their European neighbours.
This paints a really bad picture of the UK but is it accurate? In this blog I demonstrate how to use a Nest thermostat data to work out the heat loss from my house.
I don’t know how many of the 80,000 that tado used the data from were in Britain but we can assume from the type of product it would be people who were :
- technically able.
- willing to pay £50 per valve to replace a £13 thermostatic valve.
- conscious of their energy bills and possibly making other improvements to reduce energy.
My conclusion from this study was that it is probably a very conservative result, as these homes would probably have an EPC grade of C and above. 60% of the housing stock in the UK is grade D or below.
To see how I compare on this scale I needed to conduct my on experiment.
As I have been recording temperatures in my house using sensors for nearly 3 years. I have a wired outside temperature sensor on the back of my garage facing the house at about 2 m high so this is a fairly good indicator of immediate outside temperature.( I have noticed there can be a 1C difference between the front of the garage which is at the front of house facing west taken from my car sensor).
In the tado study they would only be able to use the data from the valves and these are typically placed in bedrooms living areas. You do not put thermostatic valve where you have your main heating temperature sensor that regulates you space heating. This is typically in the Hall and or Landing if you have a 2 zone heating home heating system. I wanted this experiment to be repeatable by as many others as possible and as I have a Nest Thermostat located in our hall I will show you how to get the data.
We don’t heat our hall up to 20 C and I suspect not many people do. The heat loss if the outside temperature is constant should be fairly linear so it should not matter if there is some latitude over the start temperature.
Setting up the experiment
- Identify the day when the outside temperature was 0 C for 5 hours.
- Identify the time slot where there is no use of space heating and other activity that would skew the data.
Not everyone has an outside temperature monitor so selecting a date you need to use an historical source.
If you are not keeping an eye on the weather then you could find the data from the closest met office weather station from
In our house the space heating will not come on between 12 midnight and 6:30 am. The best 5 house period is therefore 00:00 to 05:00
With Nest you can now get your data from google take-away.
Login to your google account and select manage my account select Data & privacy
Scroll down to “Data from apps and services that you use” and select “download or delete your data”
Select “Download your data”
From this screen you will see a list of “Products” click Deselect all” then scroll down to Nest
Select the check box for Nest
Then select the 3rd button “[n] types selected”
This will show a popup click “Select all” which will change to “Deselect all” click and scroll to thermostats and check. Then click OK on the popup.
Now scroll to the bottom of the screen
Click Next Step
Choose the Export Once and click “Create export”
Once this is done the screen will scroll to the top of the screen with the manage exports button but it should be quick at creating the export and automatically redirect manage exports.
Down load the report.
Save to the desk top and then extract to the contents.
Navigate into the folder select the year and the day date
Open up the csv file in excel. I like to save the file as an excel so that the formatting and selection is saved.
Freeze the view on the top row, and then set data so that we can select the day and times.
Take the first average temperature from the last and that is the heat loss.
It is very unusual to get a constant temperature over 5 hours and my out side temperature data demonstrates the subtle variations.
Wind speed and humidity may well have assisted in the heat transfer but the overall score is reasonable accurate. I don’t have an EPC grade but have used the software used by EPC assessors and calculate we are a low C.
Thanks to Andrew Handley for showing me how to use Google takeaway.