How's map-based travel estimation adjusted?
- Travel estimates provided by the map service with some changes
- Airline distance-based estimation
Travel Estimates Provided by the Map Service with Changes
The time reported by mobile workers between the end of an activity and the start of the next activity is often higher than the travel estimations provided by map services. This might be due to reasons such as unfamiliar routes, the use of longer routes for trucks, parking time, or time spent checking into customer's premises.
To improve accuracy, the application adjusts the map-based travel estimations using the historical differences between the durations reported by mobile workers and those calculated by the map service. This change applies to all map-based travel estimations, including routing using Street-Level Routing (SLR), real-time traffic, and point-to-point estimations using location coordinates.
For example, a telecommunications provider operating in a large city can ensure that mobile workers have enough time to find parking and reach high-rise customer locations, reducing late arrivals and missed SLAs. Likewise, utility or energy companies operating in metropolitan areas often dispatch mobile workers to perform maintenance or inspection tasks, where parking near the service point is limited. After moving the vehicle, a part of the time between activities may involve finding a parking spot and walking to the asset or customer site. The application leverages historical travel patterns, including cases where these short transitions consistently extend total travel time, to produce more exact, data-driven travel estimations across all operations. The result is smoother routing, fairer workload distribution, and better utilization of field resources.
How the Change Works
The adjusted travel time is calculated based on the typical difference between the map provider's estimated travel duration and the actual time between the start of an activity and the end of the previous activity reported by mobile workers. This calculation is performed separately for each travel key, provided there's enough data for those keys. The destination's travel key is used for these calculations. When there’s insufficient data for a given key, the application automatically applies the company level changes for that travel key.
There’s a new checkbox on the Configuration > Statistics page, called Adjust travels estimated by map service. Travels estimated by map services are adjusted only if this checkbox is selected. Users can set the limits to any value between 0 and 20 minutes.
If selected, a range field called Min and max changes to travel time estimates is displayed. The change based on machine learning, which is added to travel durations estimated by map providers, is always limited to this range.
For example, suppose the application learns that when travelling to a particular location (travel key), the mobile worker typically reports nine minutes more than what's estimated by the map provider. If the min and max adjustment configuration is set to 0 and 5 minutes respectively, the application adds only 5 minutes (instead of 9) to the travel time estimated by the map provider. So, if the map provider estimates 40 minutes of travel, the application schedules the activity 45 minutes after the end of the previous activity (assuming no idle time).
If the application doesn't have enough data to calculate the value of the change to be applied to SLR based travel estimations, the calculated value remains as 0. In such cases, if customers want to add some fixed value, say 10 mins to all SLR based travel durations, customers can set the Min adjustment value to 10.
For more information on how travel time is estimated using real-time traffic data and Street Level Routing (SLR), see the Estimate Travel Time Using Real-Time Data and What is street level routing (SLR)? topics.
Airline Distance-Based Estimation
With the availability of street-level routing (SLR) and real-time traffic, which have provided more correct methods for travel estimation, the application relies significantly less on the airline distance-based travel estimation. When the airline distance estimation is used, the application uses parameters specific to the travel keys if enough data is present.
- SLR isn't enabled.
- There’s insufficient data to estimate airline distance-based travel at the key level.
To simplify the configuration and maintain consistency, these parameters are now managed through the Statistics APIs, rather than the Statistics page. Administrators can use the APIs to retrieve and update airline distance-based travel estimations at the company level. This works the same way that the values are currently managed at the travel key level. This change ensures a more consistent and streamlined approach across the application.
As with airline distance-based estimation at the travel key level, the default values for travel based on airline distance can be managed through Statistics APIs.
For more information on airline distance-based estimation, see the Travel Time Prediction topic.