Wait time prediction

Abstract

The wait time prediction technology determines expected wait times for businesses or other public services using a model generated based on at least historical wait times for the business. In response to a request from a user, an expected wait time for service at the business for at least one particular time period on a particular day of a week is determined using the model and provided for display. User feedback regarding the expected wait time may be requested, and used to refresh the model as new wait times and other information are collected.

1 . A method, comprising:
 receiving data related to a business, the data including at least historical wait time information;   generating a model using one or more computing devices based on at least the historical wait time information for the business;   determining an expected wait time for service at the business for at least one particular time period on a particular day of a week using the model;   receiving a request for information about a business; and   providing for display, in response to the request, the expected wait time.   
     
     
         2 . The method according to  claim 1 , wherein generating the model is additionally based on place attributes, the place attributes comprise at least one of a place category, a number of online visits, reservation information, and user review information. 
     
     
         3 . The method according to  claim 1 , wherein generating the model is additionally based on historical temporal signals collected for a predetermined look-back window for the particular time period on the particular day of the week, the historical temporal signals comprise at least one of an arrival count, a departure count, an occupancy level, durations of visits, and a number of online visits. 
     
     
         4 . The method according to  claim 3 , wherein the historical temporal signals further comprise derived temporal signals computed over at least a part of the predetermined look-back window for at least one of the historical temporal signals for the particular time period on the particular day of the week. 
     
     
         5 . The method according to  claim 4 , wherein the derived temporal signals further comprise at least one of a backward-looking lag variable, the backward-looking lag variable being one of the historical temporal signals collected for an earlier time period on the particular day of the week. 
     
     
         6 . The method according to  claim 4 , wherein the derived temporal signals further comprise at least one of a forward-looking lag variable, the forward-looking lag variable being one of the historical temporal signals collected for a later time period on the particular day of the week. 
     
     
         7 . The method according to  claim 1 , wherein the expected wait time is an upper bound expected wait time. 
     
     
         8 . The method according to  claim 1 , further comprising:
 determining whether the expected wait time is less than a predetermined significant-wait threshold; and   displaying the particular time period as a no-wait time period if the expected wait time is less than the predetermined significant-wait threshold.   
     
     
         9 . The method according to  claim 1 , further comprising:
 computing a maximum expected wait time for the particular day of the week;   determining at least one peak interval for the particular day of the week, the peak interval comprising the particular time periods having the maximum expected wait time;   displaying the maximum expected wait time; and   displaying the at least one peak interval.   
     
     
         10 . The method according to  claim 1 , wherein the expected wait time is smoothed by taking a weighted average of the expected wait time and at least one other expected wait time from a neighboring time period. 
     
     
         11 . The method according to  claim 1 , wherein the providing for display is in association with other business information indicating busyness level. 
     
     
         12 . The method according to  claim 1 , further comprising:
 requesting user feedback regarding the expected wait time;   receiving, in response to the request, the user feedback; and   updating the model based on the user feedback.   
     
     
         13 . The method according to  claim 1 , further comprising:
 refreshing the model as new wait times are received.   
     
     
         14 . The method according to  claim 1 , wherein the model is one of a regression model or a classification model. 
     
     
         15 . The method according to  claim 14 , wherein the model is a quantile regression model capable of determining the expected wait time as a value at a predetermined quantile. 
     
     
         16 . The method according to  claim 1 , wherein the model is one of a linear model, a boosting tree model, a random forest model, or a neural net model. 
     
     
         17 . A system, comprising:
 a memory; and   one or more processor in communication with the memory, the one or more processors configured to:
 receive data related to a business, the data including at least historical wait time information; 
 generate a model using one or more computing devices based on at least the historical wait time information for the business; 
 determine an expected wait time for service at the business for at least one particular time period on a particular day of a week using the model; 
 receive a request for information about a business; and 
 provide for display, in response to the request, the expected wait time. 
   
     
     
         18 . The system according to  claim 17 , wherein generating the model is additionally based on place attributes, the place attributes comprise at least one of a place category, a number of online visits, reservation information, and user review information. 
     
     
         19 . The system according to  claim 17 , wherein generating the model is additionally based on historical temporal signals collected for a predetermined look-back window for the particular time period of the day on the particular day of the week, the historical temporal signals comprise at least one of an arrival count, a departure count, an occupancy level, durations of visits, and a number of online visits. 
     
     
         20 . A computer-readable storage medium storing instructions executable by one or more processors for performing a method, comprising:
 receiving data related to a business, the data including at least historical wait time information;   generating a model using one or more computing devices based on at least the historical wait time information for the business;   determining an expected wait time for service at the business for at least one particular time period on a particular day of a week using the model;   receiving a request for information about a business; and   providing for display, in response to the request, the expected wait time.