Thermostat scheduling using user devices and machine learning

A thermostat scheduling method, comprising:

• collecting event data for a plurality of events from one or more devices, wherein the event data for each event is associated with a user action indicative of user temperature preferences for a particular space, and comprises:
• a timestamp indicating a time during a particular day when an event associated with the user action has occurred for the particular space;
• a set point temperature value for the HVAC system for the particular space;
• and
• an occupancy status indicating whether the particular space is occupied;
• populating time entries in an occupancy history log with the event data, wherein:
• the occupancy history log comprises a plurality of time entries;
• and each time entry identifies:
• the timestamp indicating the time during the particular day when the event associated with the user action has occurred for the particular space;
• the set point temperature value for an HVAC system for the particular space;
• and
• the occupancy status indicating whether the particular space is occupied;
• identifying blank time entries among the plurality of time entries in the occupancy history log, wherein the blank time entries correspond to missing event data, sparse event data, or partially blank event data;
• populating the blank time entries by forward filling the occupancy history log using event data from another time entry occupancy history log, wherein the another time entry is for a different time on the same particular day;
• and
• outputting the populated occupancy history log.

A non-transitory computer-readable medium storing instructions that, when executed by a processor, causes the processor to:

• collect event data for a plurality of events from one or more devices, wherein the event data for each event is associated with a user action indicative of user temperature preferences for a particular space, and comprises:
• a timestamp indicating a time during a particular day when an event associated with the user action has occurred for the particular space;
• a set point temperature value for the HVAC system for the particular space;
• and
• an occupancy status indicating whether the particular space is occupied;
• populate time entries in an occupancy history log with the event data, wherein:
• the occupancy history log comprises a plurality of time entries and
• each time entry identifies:
• the timestamp indicating the time during the particular day when the event associated with the user action has occurred for the particular space;
• the set point temperature value for an HVAC system for the particular space;
• and
• the occupancy status indicating whether the particular space is occupied;
• identify blank time entries among the plurality of time entries in the occupancy history log, wherein the blank time entries correspond to missing event data, sparse event data, or partially blank event data;
• populate the blank time entries by forward filling the occupancy history log using event data from another time entry in the occupancy history log, wherein the another time entry is for a different time on the same particular day;
• and
• output the populated occupancy history log.