Instrumented People, Places and Things: aging in place, a home study, 2012 to 2013
Start here
Sensing daily life in one home. In December 2012 the Iota team at SapientNitro put sensors in the home of an elderly resident with chronic obstructive pulmonary disease (COPD). It was one person's case and a pilot for studying older people at home. Its aim, in the team's 2013 publication: “Develop a robust picture of daily living patterns within a specific elderly person’s home, in order to identify changes in behavior, or changes in bodily condition, which may suggest an emerging health problem.” The team's name for the study was Elder Care: Daily Patterns of Living in the Home.
The readings are not here. The team's viewer read them from a hosted database, which has not been found, nor have the daily files made from it. This page holds what survives to draw: the team's 3D model of the house and each instrument where the viewer saved it, drawn as the viewer drew them, with nothing to replay.
The study's anchor is a place. In TJ's words:
“If the person leaves the home they are out of the sensor, if guests visit they are also sensed. It is a place, the home. The sensors are fixed to things that stay at home.”
Moments
Click an instrument in the picture, or its row at the right, for what it read.
How to read what you see
- Orange lines
- The Kinect's field of view, as the viewer drew it: 4 m deep. It would have drawn tracked people inside it as skeletons.
- Green lines
- The Grid-EYE's field of view, down to the floor. The viewer filled its 8 × 8 cells with heat; the faint empty grid is added here.
- Pulsing cubes
- Spiders, the team's own boxes of sensors. A pulsing cube was the viewer's own sign for no reading, and with the readings gone every spider shows it. With readings it drew seven rings of sound.
- The floor
- The team's model: room names and furniture as flat gray shapes, with no photograph in it.
- Labels
- Added here, as on the open house page. Fields of view, over the picture, hides the lines.
- Placement
- The viewer saved three different placements. Over the picture you can choose among them.
What it can and cannot tell
The publication ends at early analysis. It reports no baseline of normal behavior, no activity recognized over time and no change in health. What follows is its own account, without the readings behind it.
What it reads from its graphs. With a trial set, humidity spikes in the bathroom as showers (pp. 16, 21). In the bedroom, light that tells a lamp, a ceiling light and daylight apart (p. 20). In the foyer, a person's path through the heat, and a radiator as a fixed spot (p. 20).
What it leaves open. Whether a sound in the bathroom was someone using it or other noise, and whether the lights had been left on (p. 21). Feeding, from sound and temperature, is something “we hope to be able to determine” (p. 15). Where someone was going, and why, it says “we may assume” (p. 20).
What held it back. XBee radios “could not consistently push data more than 5 feet”. Bluetooth reached farther, but its connections had to be checked daily by remote desktop (p. 16).
Where the sources differ. The publication's words on p. 15 and the viewer's graphs in the film put humidity by the living room entry and temperature in the kitchen; its p. 13, its 13 January graphs and a 2013 team deck put them the other way round. This page follows p. 15 and the viewer. In the film the Grid-EYE also sits about 20 in from its saved place; this page draws the saved one.
What survives, and what does not
| Readings | Not found: Kinect, Grid-EYE and spider, and the per-sensor daily files made from them |
|---|---|
| The model | Two parts, the slab and the walls, with baked textures, 19 December 2012 |
| Placements | Four saved files, three of them different, 19 December 2012 to 7 January 2013 |
| The viewer | Its Java source, and the firmware of its spiders, 21 December 2012 |
| The film | The viewer at work, 110 s, 10 January 2013. Only its first frame is used, on the slides |
| The publication | 33 pages, November 2013, with a day's graphs from 13 January 2013 |
| Later analysis | A script of January 2014 that smooths the bathroom spider's humidity and searches a day for the shape of one rise |
The dates are file dates. Team decks of 2013 and 2014 call this a multi-year project with over 3.2 TB and then over 6 TB of data. A folder of August 2014 sensor logs shares the study's old name; whether the home was used again, or only the name, is not known.
An Iota internal project, at SapientNitro, with no client: its sensing, sensors, hardware and software were the Iota team's. Ann Kauth worked on the project. Iota's work with sensing, this study among it, is discussed in Gaston Legorburu and Darren McColl, Storyscaping: Stop Creating Ads, Start Creating Worlds (Wiley, 2014), ch. 8, “In Their Shoes”, pp. 143–164.
This page names no one in the home and shows no reading or photograph from it. One of the studies in Instrumented People, Places and Things. A draft, November 2026.
How to move
- Rotate
- Drag, or the arrow keys
- Pan
- Right-drag, or Shift and the arrows
- Zoom
- Scroll, or Page Up and Page Down
Instruments
| Box | Instrument | Room | Read |
|---|
Box: the computer each hung from, by its number on the publication's plan (p. 14).
Model of activity
What the study set out to track (p. 8), and where the publication assigns each to a sensor (pp. 12–14). Each dot is a hypothesis about what a sensor could stand for. TJ: “An observation is not an activity model.”
Click a row to find its instruments. ADLs and IADLs are the Katz and Lawton lists the team started from (pp. 6–7). Signifiers is also a column in METIS, the team's tool for mapping a concept to what can be measured; the publication does not name METIS. Quotation: TJ, Modeling Activity, p. 3.
The eight kinds of sensor
A 2013 team deck counts “Six instrument packages, each with 8 sensors”. The publication lists eight kinds (p. 8), and its pages place six instruments with three or four sensors each, or one (pp. 13, 15–16).
Avenues for further study
The publication's list (p. 14), and phase two beside it: a Kinect in the kitchen, a power monitor on the television, and spiders with load cells under the bed and a chair. No saved placement, film or page shows these installed, and a sketch for a scale survives in the firmware.