PlaceLab Datasets

The PlaceLab is a unique live-in laboratory in Cambridge, MA. It is a joint MIT House_n and TIAX, LLC initiative. This website is intended for researchers interested in becoming more familiar with PlaceLab data and possibly using the data in their own work.

The PlaceLab was first introduced in a a CHI 2005 paper. This dataset is freely available for academic researchers to use in their own work given that this website (http://web.mit.edu/cron/group/house_n/data/PlaceLab/PlaceLab.htm) is cited whenever the dataset is used, as well this academic overview article:

S. S. Intille, K. Larson, E. Munguia Tapia, J. Beaudin, P. Kaushik, J. Nawyn, and R. Rockinson, "Using a live-in laboratory for ubiquitous computing research," in Proceedings of PERVASIVE 2006, vol. LNCS 3968, K. P. Fishkin, B. Schiele, P. Nixon, and A. Quigley, Eds. Berlin Heidelberg: Springer-Verlag, 2006, pp. 349-365. [Abstract] [PDF]

Researchers using the datasets should also indicate which of the public datasets they are using. The PLIA1 dataset is now a legacy dataset, because significant changes have since been made to the way that various data streams are shored, largely to facilitate better data sharing and visualization. The PLIA1 dataset does not work with the HandLense visualizer tool. The PLIA2 dataset uses current formats and works with the Handlense visualizer tool. We recommend that researchers interested in PlaceLab data use the PLIA2 dataset. We are leaving PLIA1 online because some work has already been published by groups outside of MIT that used it.  These two datasets are test datasets to help researchers. There is a third dataset that has been released, PLCouple1, consiting of 100 hours of annotated data from a couple who lived in the PlaceLab.

A video of the PlaceLab is also available online and provides some general information as well as shots of the interior of the facility. The citation for the video is:

S. S. Intille, K. Larson, J. Beaudin, E. Munguia Tapia, P. Kaushik, J. Nawyn, and T.J. McLeish, "The PlaceLab: a live-in laboratory for pervasive computing research (Video)," in Proceedings of Pervasive 2005 Video Program, May, 2005.

We have a new video that shows the sensor ubiquity and short latency times. It is in submission, but for an early preview, contact [intille at mit dot edu].

Some of the sensors in the PlaceLab are MITes. If data are used from them, please reference them using this publication:

E. Munguia Tapia, S. S. Intille, L. Lopez, and K. Larson, "The design of a portable kit of wireless sensors for naturalistic data collection," in Proceedings of PERVASIVE 2006, vol. LNCS 3968, K. P. Fishkin, B. Schiele, P. Nixon, and A. Quigley, Eds. Berlin Heidelberg: Springer-Verlag, 2006, pp. 117-134. [Abstract] [PDF]

In the PLCouple1 dataset, two additional sensors are included. The infrared MITes were developed as part of this work at MERL. If data are used from them, please reference them using this publication:

C. R. Wren and E. Munguia-Tapia, "Toward Scalable Activity Recognition for Sensor Networks," in Proceedings of The Second International Workshop in Location and Context-Awareness (LoCA '06), vol. 3987 / 2006, M. Hazas, J. Krumm, and T. Strang, Eds. Dublin, Ireland: Springer Berlin / Heidelberg, 2006, pp. 168-185.
[Abstract] [PDF]

The RFID tagging is provided using the Intel RFID glove. If data are used from them, please reference them using this publication:

Philipose, M., Smith, J.R., Jiang, B., Mamishev, A., Roy, S., Sundara-Rajan, K., "Battery-free wireless identification and sensing." IEEE Pervasive Computing 4(1), 37–45 (2005)

As much as our resources allow, we are operating the PlaceLab as a shared research facility. In that spirit, we are trying to make some of the data from our experiments using the PlaceLab available to other researchers interested in designing novel technologies and devices for the home. The datasets may be of particular interest to researchers in pattern recognition, since annotated multi-modal sensor datasets from real homes are challenging to obtain at this time. We have a new project that will support creation of similar types of datasets using our portable in-home sensors. Stay tuned.

Test datasets:

PLIA1 (PlaceLab Intensive Activity 1): The first test dataset we posted (and that is described in the PERVASIVE paper), is archived for those who still wish to use it. [Overview, description of data, and data]

PLIA2 (PlaceLab Intensive Activity 2): We have recently made extensive improvements to the PlaceLab infrastructure, including changing some of the formats in which data are saved to make data management and data sharing easier. These changes are reflected in the PLIA2 dataset. Researchers new to the PlaceLab should start with this test set and ignore the PLIA1 data. This dataset works with a visualization and annotation tool we have developed for PlaceLab data, called Handlense. [Overview, description of data, and data] [Overview of HandLense visualization tool and executable]

PLCouple1 (PlaceLab Couple 1): This dataset consists of all easily anonymized sensor data for a 2.5 month period when a couple stayed in the PlaceLab. Additional background information on the dataset can be gathered from these papers. :

B. Logan, J. Healey, Matthai Philipose, E. Munguia Tapia, and S. Intille, "A long-term evaluation of sensing modalities for activity recognition," in Proceedings of the International Conference on Ubiquitous Computing. vol. LNCS 4717 Berlin Heidelberg: Springer-Verlag, 2007, pp. 483–500. [Abstract] [PDF] 

J. S. Beaudin, S. S. Intille, E. Munguia Tapia, R. Rockinson, and M. E. Morris, "Context-sensitive microlearning of foreign language vocabulary on a mobile device," in Proceedings of the European Ambient Intelligence Conference 2007, 2007, p. To appear. [Abstract] [PDF] [Some interview data]

Due to subject privacy issues, it is not possible for us to release all the audio and video that was also collected for this dataset. We also have not released transcrips from all subject interviews at this time. This dataset includes readings from the Intel RFID bracelet. Annotation is costly and we have only about 100 hours of the dataset annotated at this time. More may be annotated in the future. If you can help with that effort, please let us know. [Data]

Other datasets:

We have collected other datasets from the PlaceLab and will continue to do so. We hope to make as many of those public as our resources allow. With the caveat that data that we share must be properly anonymized, we are interested in having as many other researchers benefit from the facility as possible.

If you review any of these datasets and you are interested in possibly using other datasets or working with MIT House_n researchers to conduct an original study in the PlaceLab (perhaps adding/testing new capabilities), then contact Stephen Intille ([lastname]@mit.edu). Further, contact Stephen if you have comments and suggestions on these test datasets, such as information that should be added to the documentation to make PlaceLab data easier to use, or if you have any resources you could volunteer (e.g., disk space) to make data sharing of larger datasets easier.

Disclaimer:
At this time we do not have funding explicitly for providing these datasets as a community resource. We do it by scraping by and volunteering effort. Therefore, the support we can offer is quite limited. We hope at some time in the future to obtain funding that might permit better documentation, support, etc.

Acknowledgements:

Some of the technology in the PlaceLab and some PlaceLab datasets were made possible by the National Science Foundation under Grant No.0313065. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation (NSF) .