Predictive alarm analytics
Systems and techniques are provided for identifying patterns activity at a facility that seem to precede alarm events. In some implementations, a monitoring system is configured to monitor a property and includes a sensor configured to generate sensor data. A monitor control unit is configured to receive the sensor data and based on the sensor data, determine an alarm status of the monitoring system. A monitoring application server is configured to receive the sensor data and data indicating the alarm status of the monitoring system. The server applies the sensor data to a model that determines an estimated alarm status of the monitoring system and determines the estimated alarm status of the monitoring system is different from the alarm status. Based on determining that the estimated alarm status of the monitoring system is different from the alarm status of the monitoring system, the server overrides the alarm status.
1. A monitoring system that is configured to monitor a property, the monitoring system comprising:
a sensor that is located at the property and that is configured to generate sensor data;
a monitoring application server that is configured to:
receive the sensor data;
receive geographic data for a user of the monitoring system;
determine past user behavior data for the user of the monitoring system;
determine current user behavior data for the user of the monitoring system;
apply the sensor data, the geographic data, the past user behavior data, and the current user behavior data to a model that determines an estimated alarm status of the monitoring system; and
based on the estimated alarm status of the monitoring system, control the monitoring system.
2. The system of claim 1 , wherein the monitoring application server is configured to:
based on applying the sensor data, the geographic data, the past user behavior data, and the current user behavior data to the model that determines the estimated alarm status of the monitoring system, determine a risk score of the monitoring system, the risk score indicating a likelihood that the monitoring system is at risk for an alarm;
compare the risk score to an active alarm threshold; and
based on comparing the risk score to the active alarm threshold, determine that the alarm score satisfies the active alarm threshold;
based on determining that the risk score satisfies the active alarm threshold, determine that the estimated alarm status of the monitoring system is an estimated active alarm.
3. The system of claim 2 , wherein the monitoring application server is configured to:
based on applying the sensor data, the geographic data, the past user behavior data, and the current user behavior data to the model, identify one or more resident actions to reduce the risk score; and
provide, for output, data identifying the one or more actions.
4. The system of claim 2 , wherein the monitoring application server is configured to:
based on applying the sensor data, the geographic data, the past user behavior data, and the current user behavior data to the model, identify one or more installer actions to reduce the risk score; and
provide, for output, data identifying the one or more actions.
5. The system of claim 1 , wherein the monitoring application server is configured to:
based on applying the sensor data, the geographic data, the past user behavior data, and the current user behavior data to the model that determines the estimated alarm status of the monitoring system, determine a risk score of the monitoring system, the risk score indicating a likelihood that the monitoring system is at risk for an alarm;
compare the risk score to an inactive alarm threshold;
based on comparing the risk score to the inactive alarm threshold, determine that the risk score satisfies the inactive alarm threshold; and
based on determining that the risk score satisfies the inactive alarm threshold, determine that the estimated alarm status of the monitoring system is an estimated inactive alarm.
6. The system of claim 1 , wherein the monitoring application server is configured to:
receive data indicating an actual alarm status of the monitoring system;
determine that the estimated alarm status of the monitoring system is different from the actual alarm status of the monitoring system; and
control the monitoring system by overriding the actual alarm status of the monitoring system based on determining that the estimated alarm status of the monitoring system is different from the actual alarm status of the monitoring system.
7. The system of claim 6 , wherein the monitoring application server is configured to:
receive, from the user, data verifying the overridden alarm status of the monitoring system; and
based on receiving, from the user, data verifying the overridden alarm status of the monitoring system, update the model that determines the estimated alarm status of the monitoring system.
8. The system of claim 1 , wherein the monitoring application server is configured to:
receive additional sensor data and data indicating an additional actual alarm status of the monitoring system;
apply the additional sensor data to the model to determine an additional estimated alarm status of the monitoring system;
determine that the additional estimated alarm status of the monitoring system is the same as the additional actual alarm status of the monitoring system; and
based on determining that the additional estimated alarm status of the monitoring system is the same as the additional actual alarm status of the monitoring system, maintain the additional actual alarm status of the monitoring system.
9. The system of claim 1 , wherein the monitoring application server is configured to:
train, using machine learning, the model that determines the estimated alarm status of the monitoring system.
10. The system of claim 1 , wherein the monitoring application server is configured to:
receive account status data for the user of the monitoring system; and
apply the sensor data, the geographic data, the past user behavior data, and the current user behavior data to the model that determines the estimated alarm status of the monitoring system by applying the sensor data, the geographic data, the past user behavior data, the current user behavior data, and the account status data to the model that determines the estimated alarm status of the monitoring system.
11. A method comprising:
receiving, by a monitoring application server, sensor data captured by a sensor that is located at a property monitored by a monitoring system;
receiving, by the monitoring application server, geographic data for a user of the monitoring system;
determining, by the monitoring application server, past user behavior data for the user of the monitoring system;
determining, by the monitoring application server, current user behavior data for the user of the monitoring system;
applying, by the monitoring application server, the sensor data, the geographic data, the past user behavior data, and the current user behavior data to a model that determines an estimated alarm status of the monitoring system; and
based on the estimated alarm status of the monitoring system, controlling, by the monitoring application server, the monitoring system.
12. The method of claim 11 , further comprising:
based on applying the sensor data, the geographic data, the past user behavior data, and the current user behavior data to the model that determines the estimated alarm status of the monitoring system, determining a risk score of the monitoring system, the risk score indicating a likelihood that the monitoring system is at risk for an alarm;
comparing the risk score to an active alarm threshold;
based on comparing the risk score to the active alarm threshold, determining that the alarm score satisfies the active alarm threshold; and
based on determining that the risk score satisfies the active alarm threshold, determining that the estimated alarm status of the monitoring system is an estimated active alarm.
13. The method of claim 12 , further comprising:
based on applying the sensor data, the geographic data, the past user behavior data, and the current user behavior data to the model, identifying one or more resident actions to reduce the risk score; and
providing, for output, data identifying the one or more actions.
14. The method of claim 12 , further comprising:
based on applying the sensor data, the geographic data, the past user behavior data, and the current user behavior data to the model, identifying one or more installer actions to reduce the risk score; and
providing, for output, data identifying the one or more actions.
15. The method of claim 11 , further comprising:
based on applying the sensor data, the geographic data, the past user behavior data, and the current user behavior data to the model that determines the estimated alarm status of the monitoring system, determining a risk score of the monitoring system, the risk score indicating a likelihood that the monitoring system is at risk for an alarm;
comparing the risk score to an inactive alarm threshold;
based on comparing the risk score to the inactive alarm threshold, determining that the risk score satisfies the inactive alarm threshold; and
based on determining that the risk score satisfies the inactive alarm threshold, determining that the estimated alarm status of the monitoring system is an estimated inactive alarm.
16. The method of claim 11 , further comprising:
receiving data indicating an actual alarm status of the monitoring system; and
determining that the estimated alarm status of the monitoring system is different from the actual alarm status of the monitoring system,
wherein controlling the monitoring system comprises overriding the actual alarm status of the monitoring system based on determining that the estimated alarm status of the monitoring system is different from the actual alarm status of the monitoring system.
17. The method of claim 16 , further comprising:
receiving, from the user, data verifying the overridden alarm status of the monitoring system; and
based on receiving, from the user, data verifying the overridden alarm status of the monitoring system, updating the model that determines the estimated alarm status of the monitoring system.
18. The method of claim 11 , further comprising:
receiving additional sensor data and data indicating an additional actual alarm status of the monitoring system;
applying the additional sensor data to the model to determine an additional estimated alarm status of the monitoring system;
determining that the additional estimated alarm status of the monitoring system is the same as the additional actual alarm status of the monitoring system; and
based on determining that the additional estimated alarm status of the monitoring system is the same as the additional actual alarm status of the monitoring system, maintaining the additional actual alarm status of the monitoring system.
19. The method of claim 11 , further comprising training, using machine learning, the model that determines the estimated alarm status of the monitoring system.
20. The method of claim 11 , further comprising receiving account status data for the user of the monitoring system;
wherein applying the sensor data, the geographic data, the past user behavior data, and the current user behavior data to the model that determines the estimated alarm status of the monitoring system comprises applying the sensor data, the geographic data, the past user behavior data, the current user behavior data, and the account status data to the model that determines the estimated alarm status of the monitoring system.