Constructing a reasonable mesh for road traffic crashes

[Hossein Baghishani]

Hi all,

Would it be possible to create a good mesh for the data (shown in the attached image)? If so, what should the arguments be set?

For a further explanation: the data are related to road traffic crashes where their corresponding locations are shown in the figure. For each car crash, some features are recorded.

Thanks in advance and all the best,

Hossein

[Finn Lindgren]

Hi Hossein,

assuming that you want a 2D spatial model across the space as a whole, and not a road-network restricted model, I would suggest one of these two approaches:

1. Use the crash locations as location input to inla.mesh.2d, specify "cutoff" to identify the desired resolution near the roads, and use a larger max.edge value for how densely to fill in the gaps.

2. Use inla.nonconvex.hull on the crash locations to create a "buffer zone" around the crashes/roads and use that as the inner boundary for the mesh, and specify max.edge to control the triangle size inside that subdomain and the domain outside.

Finn

--
You received this message because you are subscribed to the Google Groups "R-inla discussion group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to r-inla-discussion...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/r-inla-discussion-group/0736d020-f950-4cf0-8d9c-591385f6d5a0n%40googlegroups.com.

--

Finn Lindgren
email: finn.l...@gmail.com

[Hossein Baghishani]

Thank you so much Finn,

I'll try to work on both of your suggestions. As you noticed, I'm going to model the whole space; however, by your comment, another question engages my mind: Is there a road-network restricted model that I could implement?

Further, I plan to apply a spatial-varying Poisson model to the data. Would the road-network restricted model be a better choice for this purpose?

Best,

Hossein

[Finn Lindgren]

People have been using and working on roadwork network Poisson point process models, but I'm not aware of any ready-out-of-the-box interface for such models.

Whether a plain 2D spatial point process model works for your case probably depends on what the goal of the analysis is.

Finn