Cluster Analysis

On the Cluster Analysis page, you can perform an analysis to find the number of clusters that best matches with your real-life data. When you have created a clustering for a certain number of clusters, this is called a scenario.

You can set Transport Costing Navigator to run multiple scenarios. Because it can take a long time, you may want to run them overnight and view the results the next day.

Important

After calculations are complete for the scenarios you specified, data is updated on all Analytics pages in the app for the scenario you select on this page. You can browse the other Analytics pages to see more details.

Enter Cluster Constraints

To perform the analysis, specify the scenarios that you want to consider in Please enter your Cluster Constraints here. You can do this by specifying Min (-imum) and Max (-imum) bounds on the constraints in the first table.

The most important constraint is the Number of Clusters. The range of this constraint determines which scenarios are run.

All the other constraints determine which calculated scenarios belong in the set of feasible scenarios.

When you change the Min or Max bound for constraints, the cell background may become yellow. This means that the constraint has become binding. You can change these constraints either before or after running scenarios. E.g. when you change the minimum on the number of clusters to 10 i.o. 1, but you did run the 5 clusters scenario, this 5 clusters scenario is not feasible according to the constraints. Therefore the constraint is binding and the background is colored to indicate this. So, a yellow background means that you are somehow restricting the scenarios.

Note

  • The numbers in this table are per cluster. E.g. Min for #Customers is the minimum number of customers in a cluster.
  • Changing the vehicle data, can result in e.g. a different average vehicle load per trip. Therefore, the last three bounds are automatically reset to a non-binding value when changing the vehicle data.

Reset Constraints

When you have changed the constraints in the first table, click Reset Constraints to set them back to their default bounds. The default bounds are determined by the app based on your data.

Run Scenarios

Specify the range of scenarios (Number of Clusters) to analyze, and click Run Scenario. Depending on the lower and upper bounds for the Number of Clusters, this can take a while. After each scenario that is calculated, the data is stored.

Important

When you want to interrupt the calculations, you can do so by clicking Cancel. The scenario that is currently being run will be interrupted, resulting in an error message. Usually you can ignore these.

Please note that if you have interrupted the scenario running, data in the app can be in an intermediate state. Therefore you need to select a scenario to load a correct state again. You can do this by selecting a scenario on the Costing Curves page, or by changing the number of clusters on the Create Clusters page.

The scenarios that were already finished will still be available for analysis. If you want to know which scenarios have been fully calculated, you can check the upper (right) bound in the slider widget on the Costing Curves page (Select from Calculated Scenarios widget).

When all scenarios are calculated, the Run Scenarios button is no longer visible. Instead, the widgets related to the feasible scenarios are displayed.

Feasible Scenarios: aggregated (using selected aggregator) over scenario clusters

All calculated scenarios that match your specified constraints (Please enter your Cluster Constraints here) are compared in the Feasible Scenarios table. Please note that the table is only shown after all scenarios that match the Number of Clusters constraints have been calculated. (If the Run Scenarios button is still there, the calculations have not been completed yet.)

By default the Max aggregator is used. You can change this aggregator in the drop-down selection box widget above the table. You can choose from {Sum, Mean, Min, Max} to show the corresponding data in the Feasible Scenarios table.

Details per cluster for selected scenario

You can select a feasible scenario (identified by number of clusters) by selecting it from the drop-down selection box widget above the table. Details for that scenario appear in the Details per Cluster table.