Independent Living Sensor Review: Where Detected Activity Changed, 12 August to 6 September 2026

Across the 26 days with usable data, detected bathroom-area activity fell from an average of about 5.7 episodes per day over the first 13 days to about 3.8 per day over the last 13 days, with the lowest values in the final five days (about 2.8 per day). Daily minutes of detected bathroom-area activity moved the same way, from about 16 to about 10 minutes per day. This was not part of a general fall in activity: total detected activity averaged about 70 episodes per day with a wide daily range (47 to 106), and period comparisons give a small increase or a small decrease depending on the window chosen. The two halves of the observed period contained similar totals (936 and 893 episodes), but the bathroom area's share of all detected room activity fell from about 8% to about 5.5% while the bedroom area's share rose from about 25% to about 32%, and kitchen-area detections showed no clear change (about 37 per day in the first 13 days and 33 in the last 13, with shorter comparisons pointing the other way). The same bathroom device continued to send light and temperature readings on every observed day, including the final week.

This is the clearest change in the period and the one most worth checking against other knowledge of the individual. The direction is the same at every comparison window used, and the continued light and temperature readings from the same bathroom device make a complete sensor outage an unlikely explanation, although sensor continuity and sensitivity have not been independently verified. Fewer detections are not proof that washing or toilet use did not happen: less time in the room, a different pattern of use, or activity that did not trigger the sensor would all reduce detections. Because the overall level of activity held up and kitchen-area detections did not change clearly, the useful question is about a change in how the bathroom area is being used rather than about a general reduction in activity.

Detected bedroom-area activity rose from an average of about 17.8 episodes per day over the first 13 observed days to about 21.9 per day over the last 13 days, averaging about 27 per day in the final five days, with the highest single day (34 episodes) on 6 September. Daily minutes rose from about 64 to about 73 on the same comparison. Split by time of day, the increase sits mainly in the afternoon and evening: episodes starting between 12:00 and 17:59 rose from 88 to 118 and their detected minutes from about 318 to about 405 between the two halves, with a smaller rise between 18:00 and 21:59 and only a small change between 22:00 and 05:59. Day-to-day variation remains wide, including a low day of 44 minutes on 5 September.

The rise is in the same direction at every comparison window and is concentrated in daytime hours, so it is not explained by overnight activity. Detected bedroom-area activity records that the sensor in that area was triggered; it does not establish what the individual was doing, that they were resting or sleeping, or that more of their day was spent there. Interpretation is also limited because the lounge has no activity sensor, so an apparent increase in the bedroom cannot be weighed against time in the main living area. Read alongside the reduced bathroom-area detections, the pattern is better understood as a shift in where daytime activity is detected than as a change in how much activity there is, and is worth checking against how the individual describes their current daily routine.

Changes of area averaged about 31 per day and ranged from 11 to 66. Comparing the first 13 observed days with the last 13 gives a fall from about 33 to about 28 changes of area per day; excluding 6 September, when no hallway data exists, the same comparison gives about 33 to about 30 per day. Hallway movement and presence detections stop after 18:08 on 5 September, with none recorded on 6 September, while the same hallway device continued to send temperature and light readings throughout 6 September. Battery messages from that device became more frequent across the period (36 to 57 per day during the final week) and included a low reading of about 45% on 4 September.

The reduction in movement between areas is consistent in direction but modest, it shrinks when the final day is excluded, and it sits within the wide day-to-day variation seen throughout, so it should not be read as an established change in mobility or walking ability. The more practical point is the hallway data gap: because the same device kept reporting light and temperature while its movement channel went quiet, the missing hallway detections on 6 September are more likely to reflect a device or reporting problem than an absence of movement, and the final day's hallway and area-transition values are therefore not comparable with the other days. Checking that the hallway sensor is working, and its battery, would restore confidence in future comparisons.

Of the 36 complete local 22:00 to 06:00 windows in the requested period, only 19 contained activity evidence and could be analysed. Across those 19 nights, overnight bathroom-area activity was detected on one night only (two episodes, an average of about 0.1 per night), and no night reached three or more. The longest period without detected activity averaged about 305 minutes per night (median 296, shortest 179), and no clear change over time was identified: the direction depends on which comparison window is used.

Nothing in the overnight evidence points to frequent night-time movement or repeated bathroom-area activity on the analysed nights, which is a useful counterpoint to the daytime changes above. These figures describe detected activity and gaps between detections only: they cannot confirm sleep, waking, sleep quality or rest, and a period without detections is not evidence of sleep. Because usable evidence exists for only about half of the complete overnight windows and continuous sensor operation was not confirmed for any of them, the absence of overnight detections should not be treated as proof that nothing happened at night.

Across the 26 observed days, the first detected activity of the day ranged from 04:10 (18 August) to 10:30 (19 August), with an average of 07:19. Later first detections fell mostly between 19 and 31 August and earlier ones in the first week of September, and no consistent direction over time was identified across the comparison windows tested.

A spread of more than six hours in when the day's first detection occurs means there is no dependable 'usual' morning start in this data, which matters for planning any timed visit or support. The measure records when a sensor first registered activity, not when the individual woke, got up or began their day, and it can also shift simply because a different area was entered first. It is most useful as a prompt to ask the individual how settled or variable their mornings actually feel.

Usable data covers 26 days, from 12 August to 6 September 2026: the first 11 requested days contain no telemetry, so every comparison here describes a shorter period than requested, although all 26 observed days carry substantial activity data. Activity detection exists only for the kitchen, bedroom, hallway and bathroom areas; the lounge has no activity sensor, so time spent there is not detected at all and it appears only through an appliance power stream, which itself ends on 31 August. Leaving or returning to the property cannot be assessed because no confirmed front-door or exit sensor was detected, and exits cannot be inferred from hallway activity. Appliance-related data streams are present but usage sessions have not been derived, so nothing here evidences appliance use, including cooking. Continuous sensor operation was not confirmed for any of the overnight windows.

These gaps set the boundaries for how the findings above should be used. Room shares describe only the four monitored areas, so time spent in the lounge is invisible to this data and could account for part of what looks like a shift between rooms. Two areas that assessors often rely on, leaving the home and appliance or cooking activity, are not evidenced at all. Absence of evidence in these areas is not evidence that the activity did not occur, so questions about going out or about meal preparation need to be answered through conversation or direct observation rather than from this data.