Splitting larger clusters such as school classes into smaller groups can be an efficient measure to slow the spread of the infection. The idea is to minimize disease transmission between these smaller groups. Therefore, a splitting strategy works best if most social interaction happens between people within a group, and as few as possible between the groups (cross-cohort contacts).

If the social relationships are known, a splitting strategy along these relationship lines can prevent the spread of the infection better than a random split or a split by gender (see this study). This tool can help to find a split that minimizes cross-cohort contacts.

Instructions:

• Each student lists all classmates who are also out-of-school contacts.
• A coordinator (e.g. the teacher) collects the lists, ensures consistent spelling of names and copies the list into the input field below.
• The list has a separate line for each student, the format is "Student: Contact1, Contact2, ..."
• The distributions starts with a click of he button.

Remarks:

• Relationsships that are listed in both directions are counted twice
• To add importance to a relation, just repeat it (e.g. Adam: Ben, Ben, Ben = triple weight)
• The calculation happens locally in the browser, no data is sent to the server
• The solution is not always unique, sometimes a recalculation gives a different result
• With larger clusters the tool may not find the actual optimum but just a good aproximation (algorithm based on Kernighan/Lin)

Version: 1.01 | Contact: Andreas Neumann, Königstr. 33, D-90402 Nürnberg, Germany. email: aneumann@genics.de