Uber Technologies, Inc.

Norbert Y Hu

Aditya Somani

Neil Chagas Fernandes

Yuh-Jie Eunice Chen

Jieying Chen

Talal Toukan

Lu Huang

Pablo Bellver

Archisman Roy

Adnan Akil

Gustavo Bagdadi De Castro

Brent Evans Hamby

Calvin Shawn Lien

Brian Patrick Donovan

Systems and methods of generating active notifications for users of a networked computer system using transportation service prediction are disclosed herein. In some example embodiments, a computer system uses a prediction model to generate a transportation service prediction for a user based on an identification of the user, location data for the user, prediction time data, and historical user data for instances of the user using the transportation, and then causes a notification to be displayed on a computing device of the user based on the transportation service prediction, with the notification indicating a recommended use of the transportation service in association with the place for the time of day and the day of the week, and the notification comprising a selectable user interface element configured to enable the user to submit an electronic request for the recommended use of the transportation service.

Active notification using transportation service prediction

Various embodiments partition map data into discrete portions of map data based on predetermined boundaries, and individual map data portions (e.g., shards) are assigned to individual computing nodes in a plurality of computing nodes that service geographic routing requests. A routing component associated with the plurality of computing nodes directs the geographic routing request to one or more computing nodes. The computing nodes adaptively cache geographical map data based on, for example, data relating to past geographic routes. Once cached at a computing node in a plurality of computing nodes, the computing node services current and future geographic routing requests using its assigned portion of map data and one or more portions of map data adaptively cached at the computing node, thereby reducing chances of more than one computing node getting involved in servicing a request.

Adaptive geographical data caching based on a route pattern

Growing community of inventors

San Francisco, CA, United States of America

Norbert Y Hu