What Factors Affect Office Temperature? Comparison & Conclusion
Oct 4, 2010 8:30 AM
This is the fifth and final post in a series that shows how we conducted an observational study about office temperatures, from design to execution to analysis and visualization. See Part I, Part II, Part III and Part IV.
In the previous posts in this series, we discussed the design and execution of an observational study about what factors affect office temperature. In Parts III and IV, we saw the data displayed graphically and analytically as we explored the results. Today we will use JMP to determine what types of spaces on this floor are more likely to have a higher or lower temperature than others.
Type of space was one of the factors in our model that was determined to be significant. Here is the layout of our floor, colored by type of space so you can see what kinds of office types we’re talking about:
Common – common areas include things like break rooms and mail rooms that many people use.
Conference – the floor has two large conference rooms.
Corner – there are four large corner offices.
Exterior – the most common office type; all these offices are the same size and have windows.
Interior – these offices are on the interior of the building and have no windows.
Larger – larger offices are also window offices but get their own classification because of the different size.
Lab – there is one room on the interior of the building that fits into no other category. The lab contains numerous computers and has its own temperature settings.
How do these types of spaces differ from one another? Using Oneway Analysis in JMP, we can do comparisons for each group using Student’s t. Look at the Connecting Letters Report below, showing Level and Mean. It clearly shows that common spaces (letter A) are significantly warmer than any others. Common spaces consist of break rooms, the mail room and other spaces where multiple people are in one space at one time. In the break room and mail room, the lights are always on, and there may be multiple warm bodies in the room at the same time. This could certainly account for the higher temperatures.
Letters B and C below are very close together. Less than one degree Fahrenheit separates the highest B and the lowest C.
Letter D is the lab, in a group all alone. It’s easy to guess why – the lab has its own thermostat. There are many computers in that room, and they generate a lot of heat, so the thermostat is kept low to reduce the risk of overheating.
The final analysis I am interested in is my own office. It is an exterior office, and I’d like to know how it compares to the other exterior offices on this floor. I often get very cold in the afternoons, and I have often wondered if mine is one of the colder offices on this floor. Before starting this analysis, we’ll use the JMP Data Filter to exclude all other types of offices except exterior.
The next step is a Oneway Analysis. Because the Data Filter is in effect, only exterior offices are analyzed. I have selected the rows in the data table that correspond to my office. Dynamic linking within JMP then highlights those points on the graph. As you can see, and to my great disappointment, the temperature in my office is completely average. So much for my theory that mine was the coldest office.
JMP has allowed us to look at this data in myriad ways. This observational study was a fun way to explore many of the ways JMP can help analyze and solve problems. Because of this study, we have some good indications as to what factors are significant when it comes to predicting office temperature in our building. The results of this study could lead us to design a more robust experiment if we were further interested in modeling office temperature. Many thanks to all the volunteers who assisted with gathering data, whether you took measurements or just let us intrude into your office. It truly was a floor-wide effort.
What did you like or dislike about this series of posts about office temperature?