How to find Mean Centre of a point dataset using a Network to analyse.

[WilliamMcIntosh]

Hi All

I am wanting to create the 'Centre of Gravity' for a point dataset, weighted with the number of staff at each point. I have done this using Mean centre, however I need the mean centre tool to use the network dataset to calculate the mean centre.

I have tried Location-Allocation in Network analysis but this will not work due to the fact I don't have any facilities to chose from.

Can anyone suggest any helpful direction on this?

Thanks for your help in advance.

[DanPatterson_Retired · Accepted solution]

Could you clarify the idea of the number of staff at each point?

Are these employment centers where A has 10 employees and B has 5?

Are you are then trying to find a location along a network so that the total costs of travel to this hypothetical mean center would be minimized? (We did this for choosing a Christmas party location once.)

If so and it were a two location case, then you have A and B as origins and destination.  Find the shortest path and apportion the route based on the weight (ie A would have to travel half as far a B, where the midpoint would have a total cost of 10A and 5B, so move the midpoint along the route so that the total cost is equal.)

If the situation is multimodal, just start with the Euclidean mean center, snap to the closest road, determine the weighted costs to this point from each origin and move the point along the network until the weighted costs are equal from all origin locations... of course, this is not a press the button solution but one that could be done iteratively.

[WilliamMcIntosh]

Yes I have 62 offices with the number of staff in each. I'm wanting to find the point where a training centre would be optimally placed to have smallest travel times from all offices, however weighted by number of staff. E.g. closer to the office with 10 than the one with 5. 

[DanPatterson_Retired]

Then the starting procedure would be as I describe. (Christmas party == Training center)