Using meshgrid to analyze spatiotemporal data

Hi Everyone,

How can I create a uniform grid of a geographical area to locate and analyze the distribution of lots GPS data (vehicles trajectory data) over the time?

Any suggestion on how to use meshgrid (or any other approach) is very much appreciated.

Thank you so much

[Cris LaPierre]

Asking multiple times isn't the best approach. If you are not getting replies, it is generally because your question is too vague. Try adding additional details to your post. Better yet is to include code you have tried along with an explanation of what your code is not doing that you need help with.

Answer 1

[Ankit Dutta]

As Cris LaPierre suggested, it would be better if you could add the additional information to your post. Adding the code with explanation of what you have tried so far can be helpful to get a better understanding of the problem.

Additionally, you can try using the following functions to visualize data on maps:

• geoplot -- Plot lines in geographic coordinates
• geoscatter -- Scatter chart in geographic coordinates
• geobubble -- Visualize data values at specific geographic coordinates
• geodensityplot -- Plot for geographic density

[Navid Hashemi]

Thanks you so much for your reply, Ankit Dutta.

Cris LaPierre: Sure. From now on, I will attach the code to my questions.

[Walter Roberson]

Which coordinate system does it have to be uniform in? Uniform increments in lat / long can often be easiest to express, but it leads to cells that are not rectangles on the surface of the Earth (even ignoring altitude / terrain effects.) Rectangular in curvilinear patches approximated as flat on the surface of a deemed sphere, is not rectangular in lat / long. Rectangular in curvilenear patches approximated as flat on a reference spheroid will not necessarily agree with reference sphere...

[Navid Hashemi]

I have the latitude and longitude of the four corners and just loop through the range with incrementing 0.0003 in each step (each 0.0001 is ~10 meters). The resultant grid would have cells with an estimated 30 meter cells. I understand there are some estimations associated with it, but that is okay.

The main issue is that when I locate my data points in the created grid, one of the main tasks is to analyze the temporal variation given the times (timestamp) associated with each data point. I have not been able to find an efficient structure which let me quickly search and find the data points in the grid, given a specified range of time.