Geo-Clustering in FoodChain-Lab

Tasks

• Perform a clustering using the following workflow: https://github.com/SiLeBAT/BfROpenLa…
• Cluster all French primary producers by using the GIS Cluster node.
• Use a Max Neighborhood Distance of 100km.
• That means two stations are put into the same cluster if their distance is less than 100km.

Step 1

• Import the Example Workflow from https://github.com/SiLeBAT/BfROpenLa….

Step 2

• Drag the GIS Cluster node from FoodChain-Lab in the Node Repository to the Workflow Editor.

Step 3

• Connect the output of Joiner to the input of GIS Cluster.
• Connect the output of GIS Cluster to the first input of Tracing View.
• Double click on the GIS Cluster node to open its dialog.

Step 4

• In this dialog you can set up an algorithm for geographical clustering based on latitude and longitude.
• Click on Set Filter to define which stations should be clustered.

Step 5

• You should see this dialog now.
• Press the button in the red circle to change the Property value.

Step 6

• Select “Country”.

Step 7

• Now select “FR” as Value, since we want to cluster stations in France.
• Afterwards press Add to add another condition.

Step 8

• For the new condition select “type of business” as Property and “Primary Producer” as Value, since we want to cluster primary producers only.
• Now press OK.

Step 9

• Set the Max Neighborhood Distance to 100km. That means that stations with distance of less than 100km are put into the same cluster. For details on the algorithm look here: https://en.wikipedia.org/wiki/DBSCAN
• Press OK.

Step 10

• Right click on GIS Cluster to open its context menu and select Execute to execute the node.
• The results of the clustering are put into the new column ClusterID. This column will now be used in the Tracing View.

Step 11

• Open the Tracing View by double-clicking on it.

Step 12

• A window showing the delivery network should open now.

Step 13

• Right click in the graph to open the context menu and select Collapse by Property.

Step 14

• The clustering will be done based on the results of the GIS Cluster node.
• Select ClusterID and press OK.

Step 15

• Just press OK, since we do not want to exclude any area.

Step 16

• All French primary producers have been clustered to areas.
• Each selected station (blue circle) is an area in France.

Step 17

• Select “Picking” as Editing Mode and click in the graph to deselect all stations.
• You can now see, that one of the stations is yellow. That means, that this stations (French area) is connected to all outbreak spots (red circles).
• Press Switch to GIS to see where this area is.

Step 18

• The area is in Southern France.