Self-configuring building management system

1. A method for generating a graph data structure, comprising:

receiving, by one or more processors, data associated with a building;

generating, by the one or more processors, based on the data, one or more space nodes in the graph data structure, the one or more space nodes corresponding to respective spaces within the building;

generating, by the one or more processors, based on the data, one or more asset nodes in the graph data structure, the one or more asset nodes corresponding to respective assets within the building;

generating, by the one or more processors, one or more sensor data nodes based on sensor data and one or more edges in the graph data structure that link the one or more sensor data nodes with respective asset nodes of the one or more asset nodes;

classifying, by the one or more processors, based on the data, the one or more sensor data nodes at least by:

(i) executing, by the one or more processors, a neural network trained to extract semantic elements corresponding to the one or more sensor data nodes from the data, the neural network receiving at least one string from the data as input, and

(ii) determining classifications for the one or more sensor data nodes based on the semantic elements extracted using the neural network, the classifications used to generate relationships in the graph data structure between the one or more sensor data nodes and one or more class entities of the graph data structure corresponding to the classifications of the one or more sensor data nodes; and

generating, by the one or more processors, using the classifications of the one or more sensor data nodes, a relationship in the graph data structure between the one or more space nodes and the one or more sensor data nodes.

2. The method of claim 1 , wherein the data includes building information model (BIM) data.

3. The method of claim 1 , wherein generating the one or more space nodes includes extracting an identifier from the data associated with the building and storing the identifier with at least one of the one or more space nodes.

4. The method of claim 1 , wherein generating the one or more asset nodes includes extracting an identifier from the data associated with the building and storing the identifier with at least one of the one or more asset nodes.

5. The method of claim 1 , wherein generating the one or more edges that link the one or more sensor data nodes with the respective asset nodes of the one or more asset nodes includes dynamically controlling an environmental variable of the building, monitoring sensor measurements, and generating the one or more edges based on the monitoring.

6. The method of claim 1 , wherein classifying the one or more sensor data nodes includes assigning a tag to a sensor data node of the one or more sensor data nodes, wherein the tag is associated with at least one of a source of the sensor data represented by the sensor data node or a purpose of the sensor data represented by the sensor data node.

7. The method of claim 1 , wherein generating the relationship includes parsing the at least one string to identify a semantic representation associating a space node of the one or more space nodes and a value of the one or more sensor data nodes.

8. The method of claim 1 , wherein generating the one or more edges that link the one or more sensor data nodes with the respective asset nodes of the one or more asset nodes is further based on a source of the sensor data.

9. One or more non-transitory computer-readable storage media having instructions stored thereon that, when executed by one or more processors, cause the one or more processors to:

receive data associated with a building;

generate, based on the data, one or more space nodes in a graph data structure, the one or more space nodes corresponding to respective spaces within the building;

generate, based on the data, one or more asset nodes in the graph data structure, the one or more asset nodes corresponding to respective assets within the building;

generate one or more sensor data nodes based on sensor data;

classify, based on the data, the one or more sensor data nodes at least by:

(i) executing a neural network trained to extract semantic elements corresponding to the one or more sensor data nodes from the data, the neural network receiving at least one string from the data as input, and

(ii) determining classifications for the one or more sensor data nodes based on the semantic elements extracted using the neural network, the classifications used to generate relationships in the graph data structure between the one or more sensor data nodes and one or more class entities of the graph data structure corresponding to the classifications of the one or more sensor data nodes; and

generate, using the classifications of the one or more sensor data nodes, relationships between the one or more space nodes, the one or more asset nodes, and the one or more sensor data nodes.

10. The one or more non-transitory computer-readable storage media of claim 9 , wherein the data includes building information model (BIM) data.

11. The one or more non-transitory computer-readable storage media of claim 9 , wherein generating the one or more space nodes includes extracting an identifier from the data associated with the building and storing the identifier with at least one of the one or more space nodes.

12. The one or more non-transitory computer-readable storage media of claim 9 , wherein generating the one or more asset nodes includes extracting an identifier from the data associated with the building and storing the identifier with at least one of the one or more asset nodes.

13. The one or more non-transitory computer-readable storage media of claim 9 , wherein the instructions further cause the one or more processors to generate one or more edges that link the one or more sensor data nodes with respective asset nodes of the one or more asset nodes based on dynamically controlling an environmental variable of the building, monitoring sensor measurements, and generating the one or more edges based on the monitoring.

14. The one or more non-transitory computer-readable storage media of claim 9 , wherein classifying the one or more sensor data nodes includes assigning a tag to a sensor data node of the one or more sensor data nodes, wherein the tag is associated with at least one of a source of the sensor data represented by the sensor data node or a purpose of the sensor data represented by the sensor data node.

15. The one or more non-transitory computer-readable storage media of claim 9 , wherein generating the relationships includes parsing the at least one string to identify a semantic representation associating a space node of the one or more space nodes and a value of the one or more sensor data nodes.

16. The one or more non-transitory computer-readable storage media of claim 9 , wherein the instructions further cause the one or more processors to generate one or more edges that link the one or more sensor data nodes with respective asset nodes of the one or more asset nodes based on a source of the sensor data.

17. A system for generating a graph data structure comprising a processing circuit including a processor and memory, the memory having instructions stored thereon that, when executed by the processor, cause the processing circuit to:

receive data associated with a building;

generate, based on the data, one or more space nodes in the graph data structure, the one or more space nodes corresponding to respective spaces within the building;

generate, based on the data, one or more asset nodes in the graph data structure, the one or more asset nodes corresponding to respective assets within the building;

generate one or more sensor data nodes based on sensor data and one or more edges in the graph data structure that link the one or more sensor data nodes with respective asset nodes of the one or more asset nodes;

classify, based on the data, the one or more sensor data nodes at least by:

(i) executing a neural network trained to extract semantic elements corresponding to the one or more sensor data nodes from the data, the neural network receiving at least one string from the data as input, and

(ii) determining classifications for the one or more sensor data nodes based on the semantic elements extracted using the neural network, the classifications used to generate relationships in the graph data structure between the one or more sensor data nodes and one or more class entities of the graph data structure corresponding to the classifications of the one or more sensor data nodes; and

generate, using the classifications of the one or more sensor data nodes, a relationship between at least two of the one or more space nodes and the one or more sensor data nodes.

18. The system of claim 17 , wherein the data includes building information model (BIM) data.

19. The system of claim 17 , wherein generating the one or more space nodes includes extracting an identifier from the data associated with the building and storing the identifier with at least one of the one or more space nodes.

20. The system of claim 17 , wherein classifying the one or more sensor data nodes includes assigning a tag to a sensor data node of the one or more sensor data nodes, wherein the tag is associated with at least one of a source of the sensor data represented by the sensor data node or a purpose of the sensor data represented by the sensor data node.