Logistics
Use Cases

Route Completion Outliers

Trying to understand how delivery drivers travel from stop to stop? Use a Cascade workflow to quickly identify delivery trips that deviate from the mean for any set of routes. Using a set of bike-share data with trip level details, we can identify regular routes and outliers in just a few steps.

Let's go from messy broken up data to outliers list in just a few steps:

  • First, all of our data is broken out by month, so we can append tables together to have a broader picture
  • Calculate mean and standard deviation by start_station and end_station
  • Filter out trips that are outside of 2 standard deviations from the mean as your Outliers
  • Look at which start (or end) stations have the highest % of outlier trips

Outputs

Now that we have our outliers, we can publish this new dataset to a public URL to share with teammates so that anyone can take the reins and build out a further analysis on what might be unique about these outlier trips.

(Charts and tables are live embeds of assets produced in Cascade)

For example, in the chart below, we can see which starting stations have the highest percent of outlier trips.

(Charts and tables are live embeds of assets produced in Cascade)