Home Index About Next Previous Comments

A study of Food Consumption by Carnegie Mellon’s Computer Science department (or: observations of Bagel decay) Eric Daimler Abstract    Among the many uses for Web cams, one of them is for the detection of free food. It has been said that the intersection of under-funded graduate students and over-worked, under- socialized computer scientists is the kitchen counter at which such free food is left. In this study, we attempt to analyze the rate at which free food placed on such a counter is consumed among a population of graduate student computer scientists. We report on the behavior of such people to the presence of various types of free food. In addition to food presented in the natural course of events, we lay out sliced bagels at various times and various amounts over a one week period capturing consumption rates at various time intervals. We were able to find a correlation in the consumption of bagels in our study. Even in the presence of covariates, it appears that the log of the percentage consumed remains the same for bagels initially presented within a forty-five minutes window beginning at 10:00am. Introduction    From a cursory web search, it appears that the CSD is not alone among CS departments in its use of web cams to monitor food in some way. Beyond its humorous effect, it does serve a useful purpose: 1) extra food from that day’s meetings appears to not be wasted; 2) graduate students, already on tight budgets get regularly occurring benefit in free food (that might even be of better quality than what they were otherwise having). Among the downsides is that it might contribute to poor health by encouraging food consumption especially among heavy computer users. Methods    We purchased freshly made bagels and presented them in the CSD lounge within thirty minutes of purchase . Once selecting bagels, we chose to slice the bagels and provide varieties of bagels. We chose a local supplier, The Pittsburgh Bagel Factory (5885 Forbes Avenue, Pittsburgh, PA). The first salesperson at The Bagel Factory told us that the most popular Bagel was the sesame and the least popular was the Garlic. In addition to these two varieties, we added plain bagels as a control. Every morning, we placed the bagels in the clear plastic lid of a catering dish on the table in the CSD lounge in plain sight of the webcam. Method: At each time interval, we attempted to count the decease in overall consumption in addition to each variety. This was done with physical observation of the remaining slices. Timing: We modified our observation timing during the experiment based on the results observed. The last three experiments involved observation at fifteen-minute intervals. Results    Although we see a result with a statistically significant result, there remains some difficulty in the accuracy of this data for several reasons. Among them is data collection. Observing the bagel pieces accurately would have been desirable, but not possible unless an overhead camera was constructed. Further, our original idea of tracking bagel type proved to be impossible. While visually consumption appeared more less the same, the garlic and sesame fell off and attached it self to the other bagels in such a way that distinguishing them was nearly impossible. Further research might include some easily to distinguish varieties. Future studies should keep the following in mind when designing experiments: i) Counting progress: Different bagel types should be placed in separate trays; ii) Observation: Our researcher was caught on Tuesday and questioned about ‘the purpose’ of the Bagels. Webcam-only observation is desirable; iii) Varieties: Changing foods might be interesting, especially at different times of the day. Presenting bagels, for example, with and without spread may give its own rewarding results; iv) Slicing: Slicing presents tops and sometimes less desirable bottoms. Bottoms should probably be thrown out if too thin; v) Presentation: We presented for functionality, but controlling for beauty of the presentation could be its own experiment. Even more interesting would be a presentation of food with a rotating sign in front such as "please take only one" or the mischievous "please do not touch". Privacy    Food consumption is in the domain of human behaviors that tend not to be performed while others are actively watching. The degree to which these behaviors change depending upon explicit and/or implicit observation (both present in this experiment) is an interesting observation next to behaviors that may not change in the presence of observation. References    Camera: https://zimbs.srv.cs.cmu.edu/coke/ffc.html Project Paper (PDF)