Personal Comfort

From Personal Science Wiki
Jump to navigation Jump to search
Project Infobox Question-icon.png
Self researcher(s) Stefano Schiavon
Related tools temperature sensors
Related topics Heart rate, Environment, Temperature, Satisfaction of room temperature

Builds on project(s)
Has inspired Projects (0)
Show and Tell Talk Infobox
Featured image Personal-comfort.jpg
Date 2017/01/26
Event name Bay Area Meetup
UI icon information.png This content was automatically imported. See here how to improve it if any information is missing or out outdated.

Personal Comfort is a Show & Tell talk by Stefano Schiavon that has been imported from the Quantified Self Show & Tell library.The talk was given on 2017/01/26 and is about Heart rate, Environment, Temperature, and Satisfaction of room temperature.

Description[edit | edit source]

A description of this project as introduced by Quantified Self follows:

Stefano Schiavon is an assistant professor and researcher interested in sustainable building design. As he told us at last month’s Quantified Self meetup in Berkeley, California, “I am Italian. I love architecture. And I think buildings are beautiful.” Stefano and his colleagues have embarked on a series of studies to better understand people’s individual preferences for their environments and they are doing it with QS tools. Watch to learn some of their findings.

Video and transcript[edit | edit source]

A transcript of this talk is below:

Personal Comfort

Stefano Schiavon Just to mention there are a few places where I found Stefano’s accent a bit hard to transcribe - Alex

We are entering a time of alternative facts and I want to give sorry facts. Climate change is real and it’s human driven and we need to do something to solve it. And if we want to solve climate change, we need to think about our energy system, how we generate and how we use energy. When I was a student in an energy engineering program, I was shocked by a graph very similar to the on that you see here. Most of our gas emissions are connected to the building that we occupy. I am Italian. I love architecture and I think that buildings are beautiful and never thought a building could be responsible for such a large impact on the environment. In most of this energy in a building is using for heating, for cooling, for providing lighting, for providing good air quality. And so I thought if we were giving so much energy and resources for that we should have great space, great indoor environments like the one here designed by me. And so when I moved to Berkeley, I started analyzing the largest database in the world on the occupants of distraction, what people think about the space that they occupy, if they think they are happy with the amount of space, the art, how satisfied are they with the lighting, with the acoustic, with the cleanliness of the space. And the results are reported here for many US buildings, mainly office space, and the go-between very happy and unhappy. And the first result that you can see there is if you focus on the red dot is that the situation is not good. People are not particularly happy and we have more than 25% of the people that are dissatisfied with the size of the space that they occupy. And if you compare this with other industry for example on general people satisfied with their computers, satisfaction for an Apple Computer could be around 90 to 95%. With a Dell it could be 80 to 85%. So we are here at levels of satisfaction similar to the one of driving a Fiat car in the 90s. But when we look at a specific, the different groups the situation is even worse. Particularly if you look down you have five issues that are particularly problematic, two connected to privacy. People are uncomfortable with people looking at what they are doing and listening to what they are doing. And two are connected with temperature and air quality. Due to the fact that most of our energy is connected with these situations I focused on that. And so, how could we increase our satisfaction with space without increasing the impact of our work, and in order to do that we need to destroy an idea. And this idea that it’s possible to create one temperature that satisfies the majority of people. In the last 70 years, people in my field went around to study people. For example, suppose they wanted to measure shoes, they measure the shoe size of every person, they calculate the average and let’s say nine and they say everybody have to wear nine. Then they are surprised that we are not happy. So, in order to destroy this idea, we need to do two things. One is to develop system and equipment that can create a personalized environment, and the other one is to develop models, and today I’m going to focus on the model’s part. One option is to use what is probably already in market and known as the Nest thermostat. You start to say what you like and after awhile by some measurement of the space environment, or the temperature, or some information where you are and what you are doing, and they try to predict the temperature. We decided to go a little bit more detailed and want to measure more parameters, environmental parameters, belong to know exactly who is this person, and the next step we tried to go to biomarkers, so measuring things in your body directly. We did one of the first started, instead of having a sensor in a space and asking people what they think, we attached the person to people and then we track them for a couple of weeks. We did this in Singapore for 16 subjects and we measured temperature, relative humidity and CO2 with a frequency of one minute per week, wee asked what they think about the space and we ask them to record information where they are, are they on mass transportation, are they at home, in the office or in the restaurant. In Singapore they spend a lot of time in restaurants and that’s why they like to stay there, and we have information about the space that they occupy, do they have the air conditioning on and so on. In this graph it reports CO2 concentration versus five minutes of data for a typical person and we would like to see values around the 1000 ppm. If you measure in the United States, we expected this rarity goes much about that value. Outdoor is 500. Here we see very high levels, levels that are dangerous, not for itself, but because when those levels are high, there is a high chance that many other pollutants are going to be very high. Moreover, when you see the spikes at 2000, 3000, 4000, there is a higher risk of airborne transmitted disease like influenza. And so here, we were very surprised from the data we have the five nights this one is the one you see in green and read. This one you see are nights that the people sleep with the open windows and here are nights where people close their windows, turn on the air conditioning and basically die (sorry, that’s the wrong thing to say.) high body, high concentration. They sleep very well but then they tend to sleep even when they are at work. The spikes over here are connected to public concentration and sometimes to the sensor. So when we look at the data we see that first of all from our original objective, predicting comfort based on the environment you can see the environment there is a lot of spread of the data, and when we started to try to predict, here we use a Gaussian process method we have an incredible low ability to predict. They are not very happy with the result. So if we just measure something close to people, we can predict something better than the thermostat, the Nest but not very, very good. So, we started the project going on attaching sensors to people. We look at the lighting level and we know whether they are indoors or outdoors, if they are sleeping or not. We measure heart rate and we attach a lot of temperature sensors around their body. Some of them are in contact with the skin and some are of the clothes of the person but not directly in contact. You see here an example of the results, how the skin temperature changes during the time. We report and people have to vote on how they are thinking and in this case using the decision tree method, we were able to increase our accuracy. And particularly, we were able to have very good accuracy on understanding. If someone wanted a change in their space, we were able to provide it correctly or warmer or cooler, so we don’t do the mistake of the person that wants to be cooler and we are eating the space. And this one could be useful if we if we rode in your car, your car knows who you are and what you want, and the same information can be transferred in your office, and the office gets the temperature that you need. Or, while you are sleeping the temperature is one of the big drivers that when you wake up during the night, because for example, if you have a blanket on you, you maybe saturate with that blanket and your body says oh-oh, you know, release the heat there is a danger you wake up and turn around and you basically pump out the heat. So that could create also better sleeping. And so, why this talk, because we are looking for people like you. It’s a kind of a hard to do group experiment, we ask you to every morning, and every time you shower to install a certain amount of the sensor. We ask you to do this for two weeks and then we ask you to put the sensor in your space measuring CO2, temperature and relative humidity and if you keep the windows open closed. We asked you also to vote at least 20 times a day about the comfort, and that’s a very quick 10 seconds test. But then we are also asking you something about sleep quality before and after going to bed. And we give you some support.

Thank you very much.

About the presenter[edit | edit source]

Stefano Schiavon gave this talk.