A time-series sequence is a set of observations indexed by time. In a digital twin, sensor measurements and state changes form a record of how a physical system behaves. Precise timestamps make the sequence useful for synchronizing the twin with the physical world, monitoring current conditions, analyzing trends, and predicting what may happen next. section 4.1.2 section 8.3
Table 4.1 An example of time-series data from a Milesight AM319 environmental sensor in my home (https://resource.milesight.com/milesight/iot/document/am319-datasheet-en.pdf)
|
Sensor ID
|
Timestamp
|
CO2
|
Humidity
|
PIR
|
PM10
|
PM25
|
Pressure
|
Temperature
|
TVOC
|
|---|---|---|---|---|---|---|---|---|---|
|
24e124710b423527
|
1749945584458
|
639
|
67
|
trigger
|
6
|
5
|
1016.6
|
16.4
|
1.2
|
|
24e124710b423527
|
1749946183884
|
636
|
66.5
|
idle
|
4
|
4
|
1016.6
|
16.4
|
1.2
|
|
24e124710b423527
|
1749946784664
|
649
|
66.5
|
idle
|
5
|
5
|
1016.6
|
16.5
|
1.2
|
|
24e124710b423527
|
1749947383975
|
646
|
66.5
|
idle
|
5
|
5
|
1016.6
|
16.5
|
1.2
|
The sensor readings show successive timestamps for one device, with environmental measurements and motion state recorded together at each observation. section 4.1.2
A sequence is not necessarily a perfectly clean, synchronized feed. Measurements may arrive late or need recalibration and quality reconciliation. Keep raw readings as append-only records to preserve the original output, and maintain a curated representation when corrections are needed. High volume, frequent updates, and frequent access to recent readings also shape storage choices: queries and aggregations need to work efficiently across time. section 4.1.2
Sequences can enter a digital twin through streaming, batch, or API integrations alongside relational, spatial, object, and derived data. Industrial control systems may generate multiple readings per second. Rather than run analytical queries against safety-critical SCADA systems, a common approach is to collect and buffer their streams, then forward a copy to a database for the twin to query. section 4.8 section 4.2.1
To inspect a sequence, plot values over time, narrow the time window, compare metrics, and mark anomalies or contextual events. Grafana provides these interactions across live and historical sources. The same observations can support a progression from viewing past behavior, to detecting anomalies, to predictive modeling. section 6.4.1 section 8.3
