Individual Travel Time Rasters for Facilities, without Merging Together into Final Travel Time Raster

[mmschmitz]

Current Behavior

Currently, it seems that in AccessMod, if a set of points are uploaded into the programme, the travel time raster is calculated for each facility. The combined set of travel time rasters, for all facilities, are then merged together in a travel time raster that denotes the travel times to the combined set of facilities.

Is there a way in AccessMod, for a given scenario of set points to not dissolve this set of rasters into one larger raster, but keep the rasters separated out into each facility's travel time, with a distinct ID/disc file created for each file? I realise that this will take up important disc space, and not be feasible for large countries. However, it can be an important use case - so people can calculate travel time to facilities, without having to do it one at a time (which is what I've been doing).

And yes -I know that there is a catchment functionality in AccessMod, but I'm not looking for Voronoi-/Theissen-esque catchment areas as much as separate travel times to each facility - so that we can calculate travel times for farther populations to health facilities (if it is specified in an appropriate survey).

What Has Been Done So Far

To mimic this behavior in AccessMod 5.6.0, I have actually run facility travel times one at a time. It takes a really long time to set up everything, though.

I have also discussed this briefly with @nicolasray (via WhatsApp). However, last I checked, this was not an option.

[fxi]

Thanks for your message.

Technically, the geographic coverage does something similar in background, but the individual travel time layers are not saved. I could add an option "Preserve individual travel time layer" with a warning on the fairly large additional storage required. The development cost of this would probably be low.

For your information, I've already added an option in the geographic coverage module to set the facility capacity according to the sum of the population covered by its catchment. There is also another option to preserve the population layer instead of incrementally subtracting it from the original population. With those two options set, the geographic coverage will produce a travel time layer for each facility that looks like what you described. However, it's not saved. Yet, with this configuration, the resulting vector catchment layer can maybe fulfill your requirement.

If I misunderstood your request, do not hesitate to share additional details.

[mmschmitz]

@fxi - yes, you've understood the request correctly.

Having a button to "preserve individual travel time layers" would be the most effective way of handling this request, as it then allows for individuals to model predicted travel time to each facility. This would be helpful in a lot of use cases, especially when researchers may be simulating travel times to particular facilities that fall under a particular category for use in any geographic linking scenarios.

Creating population-independent vector layers could be useful in estimating catchment areas as well. I see this being another feature request, as this would be a different use case altogether. I can open up another request on GitHub for this.

[mmschmitz]

Also, I believe that the analytical use case you're describing in your previous post (with the vector catchment areas) was in the User Manual.

I wasn't sure how to reproduce it if I had no access to a current population distribution (i.e., didn't have a population layer, or an idea about population capacity, for the assessed population). Therefore, it was helpful for me to think about travel time catchments in a population-independent manner (i.e., like as Voronoi/Thiessen-like polygons, where we know what facility's got the shortest travel time).