User connector based on organization graph

Abstract

A system and method for identifying a user is described. The system identifies collaboration metrics based on user interaction data of users of an application from an enterprise. The system accesses enterprise organizational data of the enterprise and identifies topic data from the user interaction data and the enterprise organizational data. The system trains a machine learning model based on the collaboration metrics, the enterprise organizational data, and the topic data.

What is claimed is: 
     
         1 . A computer-implemented method comprising:
 identifying collaboration metrics based on user interaction data of users of an application from an enterprise;   accessing enterprise organizational data of the enterprise;   identifying topic data from the user interaction data and the enterprise organizational data;   training, at a server, a machine learning model based on the collaboration metrics, the enterprise organizational data, and the topic data;   identifying, using the machine learning model, a second user as an expert corresponding to a topic attribute; and   causing a display of an identification of the second user on a first device of a first user.   
     
     
         2 . The computer-implemented method of  claim 1 , further comprising:
 identifying reply trends from expert users based on a number of degree of separation from users requesting expert assistance from the expert users,   wherein training the machine learning model is based on the reply trends.   
     
     
         3 . The computer-implemented method of  claim 1 , further comprising:
 updating, at the server, a directory application based on the machine learning model by updating the enterprise organizational data with the topic data and skill ranking data of the expert.   
     
     
         4 . The computer-implemented method of  claim 1 , wherein identifying the second user comprises:
 ranking one or more users as the expert for the topic attribute based on social network metrics between the first user and the users of the application, and profiles of the users of the application, profiles of the users comprising a reach index, an expert page rank, a collaboration time on a first topic, past requests response times, request response rates, referral recommendations, and open requests in request queues of the users.   
     
     
         5 . The computer-implemented method of  claim 1 , wherein the user interaction data include collaboration data and the topic data from a combination of emails, meeting requests, meeting responses, and instant messages associated with the users of the application. 
     
     
         6 . The computer-implemented method of  claim 5 , wherein the collaboration data identify enterprise users interactions between users of the application, the enterprise users interactions indicating how often users communicate with each other, how often the users hold meetings, how often the users collaborate on common projects, when the users communicate with each other. 
     
     
         7 . The computer-implemented method of  claim 5 , wherein the topic data identify topic attributes from a combination of emails subject header, preset keywords in a body of emails or messages, and a title of a file attached to emails,
 wherein identifying the topic data further comprises:   extracting the topic attributes from enterprise users communication; and   parsing messages for key topics related to a set of topics predefined for the enterprise.   
     
     
         8 . The computer-implemented method of  claim 1 , further comprising:
 operating, by using the machine learning model, one of an anonymized expert recommendation workflow, an influencer identification workflow, a skills marketplace workflow, or an enterprise user graph visualization workflow.   
     
     
         9 . The computer-implemented method of  claim 8 , wherein the anonymized expert recommendation workflow is configured to:
 identify, using the machine learning model, a third user as the expert for the topic attribute; and   in response to identifying the third user, provide an anonymous recommendation of the third user to the first user;   in response to identifying the third user, communicate an expert request to the third user;   in response to communicating the expert request, receive an approval response from a third device of the third user; and   in response to receiving the approval response from the third device, provide an identification of the third user to the first device.   
     
     
         10 . The computer-implemented method of  claim 8 , wherein the influencer identification workflow is configured to:
 identify an influencer user based on collaboration metrics of the influencer user exceeding a preset collaboration metric thresholds,   wherein the preset collaboration metric thresholds indicate thresholds of how often the influencer user is invited to meetings, whether the influencer user leads a team based on a profile of the influencer user, and a number of users the influencer user collaborates with.   
     
     
         11 . The computer-implemented method of  claim 8 , wherein the skills marketplace workflow is configured to:
 generating skill attribute values of users based on the machine learning model; and   identify a skillset user with specific skillsets based on the skill attribute values of the skillset user.   
     
     
         12 . The computer-implemented method of  claim 8 , wherein the enterprise user graph visualization workflow is configured to:
 cause a display of a graphical representation of a collaboration-topic graph based on the machine learning model.   
     
     
         13 . The computer-implemented method of  claim 1 , wherein the enterprise organizational data indicates a profile for each user, and a network relation graph of the users. 
     
     
         14 . The computer-implemented method of  claim 13 , wherein the profile of for each user comprises a human resource attribute, a skill attribute, and a project attribute,
 wherein the network relation graph indicates a degree of connections, a closeness index, and an influence index.   
     
     
         15 . The computer-implemented method of  claim 1 , wherein the machine learning model identifies the second user based on a network degree of connection between the first user and the second user, network metrics of the first user and the second user, topics associated with the second user, a number of emails composed by the second user where the emails indicate a first topic, and a frequency of collaboration between the second user and other users on projects related to the first topic. 
     
     
         16 . The computer-implemented method of  claim 1 , wherein the collaboration metrics for a user indicate at least one of a collaboration topic, a collaboration pattern, a collaboration time on a topic from the user, a number of past requests for the user, response times from the user, or a request-response rate of the user. 
     
     
         17 . The computer-implemented method of  claim 1 , further comprising:
 updating collaboration metrics based on updated user interaction data of the users;   accessing updated enterprise organizational data of the enterprise;   updating the identified topic data from the updated user interaction data and the updated enterprise organizational data; and   retraining the machine learning model based on the updated collaboration metrics, the updated enterprise organizational data, and the updated identified topic data.   
     
     
         18 . The computer-implemented method of  claim 17 , wherein the application comprises at least one of an email application, an instant message application, a document sharing application, a meeting application, or a calendar application. 
     
     
         19 . A server comprising:
 one or more processors; and   a memory storing instructions that, when executed by the one or more processors, configure the server to perform operations comprising:   identifying collaboration metrics based on user interaction data of users of an application from an enterprise;   accessing enterprise organizational data of the enterprise;   identifying topic data from the user interaction data and the enterprise organizational data;   training, at the server, a machine learning model based on the collaboration metrics, the enterprise organizational data, and the topic data;   identifying, using the machine learning model, a second user as an expert corresponding to a topic attribute; and   causing a display of an identification of the second user on a first device of a first user.   
     
     
         20 . A non-transitory computer-readable storage medium, the computer-readable storage medium including instructions that when executed by a computer, cause the computer to perform operations comprising:
 identifying collaboration metrics based on user interaction data of users of an application from an enterprise;   accessing enterprise organizational data of the enterprise;   identifying topic data from the user interaction data and the enterprise organizational data;   training, at the computer, a machine learning model based on the collaboration metrics, the enterprise organizational data, and the topic data;   identifying, using the machine learning model, a second user as an expert corresponding to a topic attribute; and   causing a display of an identification of the second user on a first device of a first user.