htrebilcock
commented
1 day ago Thanks Sammi - I think until we have ventured onto the geographical modelling we should stay away from a geographical animation as it's likely to confuse everyone. They will assume I think that the animation is showing the geographical modelling. I think we should go with your first suggestion of resource focused, as this is the purpose of the DES model - to better understand the utilisation of our resources and make informed decisions about future resources.
Animated versions of models can help non-specialists to grasp what is going on in the simulation.
Possible animation approaches
Simple - resource-focussed
The simplest version of the animation I can think of will just focus on the status of any resources at a given time. e.g. is each helicopter currently unavailable, idle, on a job, transporting patient, returning to base (categories can be added/removed from that list).
Simple - call-focussed
A different version could instead come from the call perspective - so for each individual call, what is happening to it at that point in the simulation?
Advanced - psuedo-geographical
The final version I can think of is a pseudo-geographical version. My understanding from previous conversations is that flight speed is pretty consistent - so if we know how long our sampled journey time is, we can make an assumption about how far it is from where the helicopter started.
(some potential complications there around instances where the helicopter goes straight from one job to another? Or additional time taken due to complex landing site or other factors - but not sure this is an issue in this visualisation as it's always going to be a simplification of reality)
Initial thoughts are the background for this animation would be a map encompassing full extent of DAA area. We convert journey time and convert that to distance (with some flex). We then try to generate a plausible coordinate point for that call to have come in from - i.e. land-based. This is then shown on the 'map' as part of the animation. Calls could be represented in visually distinct ways to show e.g. different outcomes. They potentially just stay as part of the animation until they are cleared.
My initial feeling is that this could help people further to engage with the model - however, it may be relatively complex to implement and, given that we don't have a true geographical element to the model at this stage, there is the potential for it to be fundamentally misrepresenting how the model is functioning, which could conversely reduce trust. There are also likely to be some instances of the calls being in 'unlikely' locations, which people may get hung up on unnecessarily.
(notes to self - LSOA travel times and subsequent sampling? How many possible starter locations for the helicopter resource are there? How to factor in critical care cars to this approach - calls that are outside the radius of where the cars cover in reality?)