Geographic Weighted Regression on 911 phone calls 

Amazing tutorial! I've been wanting for some time to have some knowledge on OLS and GWR and your video was the perfect introductory choice to start learning about it. Thank you so much!

Stellar tutorial and walkthrough! Thank you!

This tutorial is amazing. I've been looking for detailed examples and this came to my rescue. Thank you!

Thank you so much for sharing this super great tutorial!!!!!! learned so much from it!!!!!

Thank you for sharing! We all appreciate your hard efforts and talent for making this video :-) you have really helped with my project, thank you!

Thanks for sharing this tutorial.

Thank you very much for that amazing video lecture.

thanks for your video, It was useful and easy to follow

Great contribution :)

Thank you for this awesome video! Please do more videos about this later :)

thanks for sharing！it's really help！

Very much appreciated!

Thank you so much. It helped a lot. God bless you.

Excellent Well Done

Thank you !!

Thanks for the clear tutorial. Although, was a bit unclear towards the end where you ran another GWR to get the predicted values layer which you compared with the observed. In the previous video frames, you showed a table with the predicted values already set for the polygons. So, unclear you ran the GWR again.

Good teacher

It should work, the toolbox is a little different, but if you have the spatial analyst extensions the tools should be basically the same.

You are welcome.

I think you are referring to the random distribution of the standard residual? The reason we want the random distribution is because what the residual represents and what not having random signifies. The residual represent the difference between the predicted value from the regression and the observed value. If this residual is not randomly distributed then that mean there is some factors that cause variation not taken into account. Having random means we are representing the causes well.

Hi, I am applying this methodology in one of my papers. As far as I learned from this video, the application of GWR requires that linear regression must be tested regarding the assumptions of variables linear dependence, error normality and error constant variance. Is there any document that I can site where this procedures is demonstrated? Many thanks for this video, it helped me a lot!!!!

Hello, tks for your tutorial. I read that WSR can be used even when Moran indexes indicate spatial autocorrelation. However in your tutorial, as far I as could understand, you put it as a condition to run such analysis. Would you mind explaining it better? ks a lot

I put the URL in the description.

I also fint this strange. If you are able to fit a good OLS regression on the data with no autocorr in residuals why would you need the GWR? To my understanding GWR is supposed to be used when you suspect spatial autocorrelation to get rid of it?

Dataset link update: www.dropbox.com/s/8c2jfcs8d82emoc/Lab_GWR.zip?dl=0

Dear Moulay Based on regression coefficient can we create a pollution surface/concetration surface. would you able to show a video for that. best regards hard and excellent work Jayan

can you plz explain what the under prediction and over prediction means in the map? badly need this!!!

This error is a hard one to decode, normally it arises when there is some kind of file error, like the disk is full, you reached your network quota, of there is sharing violation with the file you are working with. My suggestion is to export the original file to a new place on the Hard Drive and run again. If that doesn't work check your scratch workspace, make sure it is not in a place on the hard drive that is near full.

Is there any such tutorial/feature for QGIS?

Hi, this is Carol and thank you for your excellent video. Now I have faced a little bit troubles when I put my variables in the GWR model in ArcGIS, so I don't know if you could give me a hint. 1. The GWR model in ArcGIS could not run with a dummy variable in, right? What about categorical variables? 2. After I run an OLS model in ArcGIS, I want to run a GWR model continuously(without dummy and categorical variables and the VIF value of all the variables are below 5). However, only 5 of 10 independent variables could run in it? I don't know what is happening with the other 5 variables? Is there any other test or limitation I don't know so it couldn't run successfully? I will be grateful if you could give me some guidance. Thank you.

It took me sometime to figure out how to create a properly formatted Unique ID field for the OLS, using some python code. Since this is such a finicky thing, perhaps a brief video with a link to it here, would be useful.

Is it possible to perform spatial regression with data that does not overlap spatially? For example, I want to relate nitrogen measurements in an estuarine water body to septic tanks, landuse, etc on land.

Excuse me... if I have point data instead of polygon in micro-area about 9km square. Is it possible to do like this ?

I can't add a unique ID field. The window does not show my unique ID field for my OLS.

I am also going to use GWR for my master's degree but still very new to GWR. I have question for data size. You used over 2000 points of 911 calls. I only have over 1000 points. Would 1000 points be able to execute GWR with good accuracy? How much data size should I use? It is really helpful for me. Thank You Sir!

Yeah this is not my typical RU-vid video. If you look at the others on my channel I try to make them short and task based. But for this assignment for my class it was easier to have a walk through of the entire lab. Normally students will follow along with the video. You can find the datasets needed at the blog: southwesterngis.blogspot.com

This video is great thanks! I do encounter a problem when I run the GWR the same way you run this in the video, I get ... Error 99998: Unexpected Error. No output rasters are produced. Can't work around this problem. Using ArcGIS 10.1 with Windows 7 Any suggestions?

Can you please share the step by step of how to prepare the data from Hot Spot analysis to GWR dataset?

please i am final year undergraduate student my project is to build a landslide susceptibility map using logistics regression. i have generated the following maps so far.. slope map, hill shade, curvature, elevation, road buffers, river buffers of 100 - 500m , Aspect, land use land cover, ndvi. all this maps after production do not have attribute table... please i need a detailed explanation on how to use some of this data produced to produce a susceptibility map using logistic regression.. Thank you for your soonest possible reply..

Thank you for sharing your knockledge with us. Dropbox links doesn´t work. Could you replace it to help us to download dataset, please?

Just want to ask if it still works on ArcGIS 9.3? tnx

how can I download this package (GWR) for Arcgis or Qgis? Thanks in advance.

hi! i got a special error on GWR!and there is no result because of it! "Results cannot be computed because of severe model design problems. " can anyone help me?

flaco haceme la otra parte

Amazing tutorial! Could I get the dataset for this tutorial? The dropbox link is not found, thank you.

Great video. Could you please share dataset with us?

Very interesting can you please share data used forn this , i want test some thing. Thanks

thanks for your video, can you please share data,I want to test some thing. Thanks

sir, please send the pdf of this tutorial. I am not able to download it. plz send in this email id.

This is quite an impressive video. How long did it take to make? Have you considered making shorter videos? I ask because this is one of the longest educational videos I've seen on RU-vid. In any case, thanks for sharing!

A HUEVO BROTHER..