Please enable Javascript
Skip to main content

Uber Health’s Availability Predictor

The Availability Predictor feature uses machine learning to estimate and display the likelihood of driver availability, so that you can request rides with confidence.

Frequently asked questions

  • The Availability Predictor feature uses machine learning to predict the likelihood of driver availability based on historical data. This is meant to give your organization greater confidence when using the Uber Health platform.

    With this new feature, you’ll no longer be blocked from scheduling rides in advance in particular areas. Our goal is to provide you with the information you need to make a decision to request a ride anywhere, at any time.

  • The predictions are made at the trip level. The machine learning model considers historical data such as the day of week, time of day, ride option type (UberX or UberXL, for instance), location of the trip, and more. The model will continue to improve and is subject to change.

    • High availability (indicated in green): Over 90% confidence in driver availability based on historical data
    • Medium availability (indicated in yellow): 70% and 90% confidence in driver availability based on historical data
    • Limited availability (indicated in red): Under 70% confidence in driver availability based on historical data
  • When requesting a ride on the Uber Health dashboard, you’ll see an availability prediction indicator that says either High availability, Medium availability, or Limited availability.

  • Because the Availability Predictor model is based on historical data and doesn’t account for real-time considerations, you will only see predictions when requesting rides for a later date up to 30 days ahead.

    For on-demand or same-day trip requests, you can instead look in the Uber Health dashboard in the “Vehicle type” menu. Use the estimated driver arrival time provided for each ride option (for example, “in 5 minutes”) as an indicator of availability. That ETA will consider current factors, such as weather conditions, traffic, and more.

  • When the Availability Predictor is turned on for your organization, Uber Health no longer blocks access to request rides in advance in areas that were previously restricted for reliability reasons. Note: There will still be a few locations that are restricted apart from reliability reasons.

    The machine learning model is a better way to predict availability than the previous restrictions, as it updates more frequently and is more specific. We recommend that you use the predictor on all scheduled rides to have a sense of availability in advance.

  • When requesting rides that show medium or limited availability, we suggest that you monitor the status of these trips closely. If the rider does not get picked up, you can try the request again or make alternative arrangements.

Predictions are estimates, based on historical data including day of the week, time of day, vehicle type, location, and more, and are subject to change.