Automatic prediction of popularity to specified points of interest

Abstract

Techniques are described herein for predicting popularity metrics and/or visitation metrics that are used in the selection of a point of interest (POI) for placement of an electric vehicle charging station (EVCS). The techniques involve training a machine learning model based on information obtained about POIs at which EVCSs are already installed. The information used to train the machine learning model includes, for each existing installation location: (a) visitation data that describes visitation features, and (b) popularity metrics and/or visitation metrics that have been generated for the location. When the machine learning model has been trained, the trained machine learning model predicts popularity metrics and/or visitation metrics for a POI location at which no EVCS has been installed based on the visitation data of that POI.

What is claimed is: 
     
         1 . A computer-implemented method comprising:
 training a machine learning model to predict at least one popularity metric that reflects popularity of points of interest (POIs) for which popularity metrics are not currently available;   wherein training the machine learning model comprises training the machine learning model using a training dataset comprising:
 visitation data for one or more POIs for which popularity metrics are currently available, and 
 popularity metrics for the one or more POIs; 
   using the trained machine learning model, determining a particular popularity metric for a particular POI for which popularity metrics are not currently available based on one or more inputs associated with the particular POI.   
     
     
         2 . The method of  claim 1 , further comprising:
 selecting the particular POI for placement of an electric vehicle charging station (EVCS) based, at least in part, on the particular popularity metric satisfying one or more criteria.   
     
     
         3 . The method of  claim 1 , wherein the particular popularity metric comprises a percentage of visit count per multiple segments of a time frame for the particular POI. 
     
     
         4 . The method of  claim 1 , wherein the one or more inputs includes one or more of: a road classification of a nearest road, a population density in a surrounding area, and a number of other POIs of the same category as the particular POI within a vicinity. 
     
     
         5 . The method of  claim 1 , furthering comprising:
 generating, based on the particular popularity metric, instructions that specify to deliver charge to electric vehicles over a specific time frame at a specific rate;   transmitting the instructions to the EVCS to cause the EVCS to deliver charge to electric vehicles over the specific time frame at the specific rate.   
     
     
         6 . The method of  claim 1 , furthering comprising:
 calculating a deficiency value and a served value for a specific type of charger, wherein the deficiency value and the served value are calculated based on an essential population of electric vehicle drivers in a same category as the particular POI that exist in an area around the particular POI and one or more charging recommendations that indicate an amount of chargers of the specific type that the area around the particular POI can support;   calculating a first partial visitation lift value based on the deficiency value;   calculating a second partial visitation lift value based on the served value;   calculating a total visitation lift value for the particular POI based on the first partial visitation lift value and the second partial visitation lift value.   
     
     
         7 . The method of  claim 6 , wherein selecting the particular POI for placement of an EVCS is additionally based on the total visitation lift value for the particular POI. 
     
     
         8 . The method of  claim 1 , wherein the one or more inputs includes traffic data for roads near the one or more POIs. 
     
     
         9 . The method of  claim 1 , wherein the visitation data includes attributes as an aggregated distribution over two-dimensional or three-dimensional regions. 
     
     
         10 . The method of  claim 1 , wherein determining the particular popularity metric for the particular POI for which popularity metrics are not currently available further comprises determining the particular popularity metric in aggregate for multiple POIs in a region that includes the particular POI. 
     
     
         11 . A system comprising:
 one or more processors;   one or more storage devices operatively coupled to the processor;   instructions, stored on the one or more storage devices, which, when executed by the one or more processors, cause:   training a machine learning model to predict at least one popularity metric that reflects popularity of points of interest (POIs) for which popularity metrics are not currently available;   wherein training the machine learning model comprises training the machine learning model using a training dataset comprising:
 visitation data for one or more POIs for which popularity metrics are currently available, and 
 popularity metrics for the one or more POIs; 
   using the trained machine learning model, determining a particular popularity metric for a particular POI for which popularity metrics are not currently available based on one or more inputs associated with the particular POI.   
     
     
         12 . The system of  claim 11 , further comprising:
 selecting the particular POI for placement of an electric vehicle charging station (EVCS) based, at least in part, on the particular popularity metric satisfying one or more criteria.   
     
     
         13 . The system of  claim 11 , wherein the particular popularity metric comprises a percentage of visit count per multiple segments of a time frame for the particular POI. 
     
     
         14 . The system of  claim 11 , wherein the one or more inputs includes one or more of: a road classification of a nearest road, a population density in a surrounding area, and a number of other POIs of the same category as the particular POI within a vicinity. 
     
     
         15 . The system of  claim 11 , wherein the instructions further comprise instructions for:
 generating, based on the particular popularity metric, instructions that specify to deliver charge to electric vehicles over a specific time frame at a specific rate;   transmitting the instructions to the EVCS to cause the EVCS to deliver charge to electric vehicles over the specific time frame at the specific rate.   
     
     
         16 . The system of  claim 11 , wherein the instructions further comprise instructions for:
 calculating a deficiency value and a served value for a specific type of charger, wherein the deficiency value and the served value are calculated based on an essential population of electric vehicle drivers in a same category as the particular POI that exist in an area around the particular POI and one or more charging recommendations that indicate an amount of chargers of the specific type that the area around the particular POI can support;   calculating a first partial visitation lift value based on the deficiency value;   calculating a second partial visitation lift value based on the served value;   calculating a total visitation lift value for the particular POI based on the first partial visitation lift value and the second partial visitation lift value.   
     
     
         17 . The system of  claim 16 , wherein selecting the particular POI for placement of an EVCS is additionally based on the total visitation lift value for the particular POI. 
     
     
         18 . The system of  claim 11 , wherein the one or more inputs includes traffic data for roads near the one or more POIs. 
     
     
         19 . The system of  claim 11 , wherein the visitation data includes attributes as an aggregated distribution over two-dimensional or three-dimensional regions. 
     
     
         20 . The system of  claim 11 , wherein determining the particular popularity metric for the particular POI for which popularity metrics are not currently available further comprises determining the particular popularity metric in aggregate for multiple POIs in a region that includes the particular POI.