Generating catalog-item recommendations based on social graph data

Abstract

In one embodiment, a method includes sending instructions for presenting one or more recommended vendors to a client system associated with a first user via an application associated with a computing system, wherein the one or more recommended vendors are selected based on location data and user data associated with the first user, receiving a user selection of a recommended vendor from the one or more recommended vendors from the client system via the application, sending instructions for presenting an ordering interface comprising one or more catalog items and one or more delivery options to the client system via the application, receiving a user selection of one or more of the catalog items and one of the delivery options from the client system via the application, and sending an order based on the user selected catalog items and delivery option to a third-party system associated with the user selected vendor.

What is claimed is: 
     
         1 . A method comprising, by a computing system:
 sending, to a client system associated with a first user via an application associated with the computing system, instructions for presenting one or more recommended vendors, wherein the one or more recommended vendors are selected based on location data and user data associated with the first user;   receiving, from the client system via the application, a user selection of a recommended vendor from the one or more recommended vendors;   sending, to the client system via the application, instructions for presenting an ordering interface comprising one or more catalog items and one or more delivery options;   receiving, from the client system via the application, a user selection of one or more of the catalog items and one of the delivery options; and   sending, to a third-party system associated with the user selected vendor, an order based on the user selected catalog items and delivery option.   
     
     
         2 . The method of  claim 1 , wherein the user preference vector corresponds to one or more stated user preferences of the first user and one or more inferred user preferences of the first user. 
     
     
         3 . The method of  claim 2 , further comprising:
 determining one or more stated user preferences of the first user and one or more inferred user preferences of the first user based on analysis of activity of the first user on an online social network.   
     
     
         4 . The method of  claim 2 , wherein one of the one or more inferred user preferences comprise a food preference, an order preference, or a vendor preference. 
     
     
         5 . The method of  claim 1 , wherein providing the first reference of the plurality of references comprises generating a recommendation regarding the first reference for the first user. 
     
     
         6 . The method of  claim 1 , further comprising:
 receiving information regarding a completed order of the first user that comprises order parameters associated with the completed order and vender parameters associated with the completed order;   determining feedback of the first user regarding the completed order based on the information regarding the completed order; and   updating the user preference vector for the first user based on the order parameters associated with the completed order, the vender parameters associated with the completed order, and the feedback of the first user regarding the completed order.   
     
     
         7 . The method of  claim 1 , wherein the providing the first reference of the plurality of references is further based on a calculation of an affinity coefficient between the first user and the first reference. 
     
     
         8 . The method of  claim 7 , further comprising:
 calculating the affinity coefficient by determining a weighted combination of a plurality of factors, at least one factor corresponding to an interaction of the first user on an online social network with the catalog item associated with the first reference or the vendor associated with the first reference.   
     
     
         9 . The method of  claim 7 , further comprising:
 calculating the affinity coefficient based on accessing a social graph on an online social network, the social graph comprising a plurality of nodes and a plurality of edges connecting the nodes, each edge representing a relationship between two nodes, wherein a first node of the plurality of nodes corresponds to the first user, wherein one or more second nodes of the plurality of nodes correspond to the catalog item associated with the first reference, the vendor offering the catalog item associated with the first reference, or a food concept node.   
     
     
         10 . The method of  claim 1 , further comprising:
 calculating a score for each reference of the plurality of references based at least in part on the one or more order parameters and the one or more metadata items; and   wherein the first reference is provided based on the score calculated for the first reference.   
     
     
         11 . The method of  claim 1 , further comprising:
 accessing, for each reference of the plurality of references, a catalog-item vector constructed based on the one or more metadata items corresponding to the reference provided by the vendor of the catalog item corresponding to the reference;   calculating a score for each reference of the plurality of references based at least in part on a difference between the user preference vector and the catalog-item vector; and   wherein the first reference is provided based on the score calculated for the first reference.   
     
     
         12 . The method of  claim 1 , wherein each vendor corresponds to a vendor node in a social graph maintained by an online social network, the social graph further comprising a plurality of user nodes corresponding to one or more users of the online social network, the social graph further comprising a plurality of edges connecting nodes, each edge representing a relationship between two nodes, the method further comprising:
 calculating a score for each reference of the plurality of references based on a number of edge connections between the vendor node corresponding to the vendor associated with the reference and one or more user nodes, wherein a higher number of edge connections corresponds to an increased score.   
     
     
         13 . The method of  claim 1 , wherein the order parameters comprise a location of the client system, the method further comprising:
 providing, to the client system of the first user, a recommendation to place an order with a first vendor of the one or more vendors through an application executing on the client system based on comparison of the location of the client system to a location of the first vendor.   
     
     
         14 . The method of  claim 1 , wherein the one or more metadata items for each reference comprise a delivery time associated with the vendor offering the catalog item associated with the reference. 
     
     
         15 . The method of  claim 1 , wherein the one or more metadata items for each reference comprise a user rating associated with the vendor offering the catalog item associated with the reference. 
     
     
         16 . One or more computer-readable non-transitory storage media embodying software that is operable when executed to:
 access one or more order parameters associated with a first user, the order parameters comprising one or more user-specified parameters;   determine a user preference vector for the first user;   access a plurality of references to catalog items offered by one or more vendors, wherein each reference is associated with one or more metadata items provided by the respective vendor; and   provide, to a client system of the first user, a first reference of the plurality of references based on at least a comparison of the one or more order parameters, the user preference vector, and the one or more metadata items corresponding to the first reference.   
     
     
         17 . The computer-readable non-transitory storage media of  claim 16 , wherein the user preference vector corresponds to one or more stated user preferences of the first user and one or more inferred user preferences of the first user. 
     
     
         18 . The computer-readable non-transitory storage media of  claim 17 , wherein the software is further operable when executed to:
 determine one or more stated user preferences of the first user and one or more inferred user preferences of the first user based on analysis of activity of the first user on an online social network.   
     
     
         19 . A system comprising: one or more processors; and a non-transitory memory coupled to the processors comprising instructions executable by the processors, the processors operable when executing the instructions to:
 access one or more order parameters associated with a first user, the order parameters comprising one or more user-specified parameters;   determine a user preference vector for the first user;   access a plurality of references to catalog items offered by one or more vendors, wherein each reference is associated with one or more metadata items provided by the respective vendor; and   provide, to a client system of the first user, a first reference of the plurality of references based on at least a comparison of the one or more order parameters, the user preference vector, and the one or more metadata items corresponding to the first reference.   
     
     
         20 . The system of  claim 19 , wherein the user preference vector corresponds to one or more stated user preferences of the first user and one or more inferred user preferences of the first user.