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!
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.
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.
Yeah this is not my typical UA-cam 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
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?
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 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!
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.
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.
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.
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?
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
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.
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..
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?
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 UA-cam. In any case, thanks for sharing!
This tutorial is amazing. I've been looking for detailed examples and this came to my rescue. Thank you!
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!
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!
Tori Johnson Happy to help!
Stellar tutorial and walkthrough! Thank you!
Thank you so much for sharing this super great tutorial!!!!!!
learned so much from it!!!!!
Dataset link update: www.dropbox.com/s/8c2jfcs8d82emoc/Lab_GWR.zip?dl=0
Thank you for this awesome video! Please do more videos about this later :)
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.
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.
Yeah this is not my typical UA-cam 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
Is there any such tutorial/feature for QGIS?
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?
can you plz explain what the under prediction and over prediction means in the map?
badly need this!!!
It should work, the toolbox is a little different, but if you have the spatial analyst extensions the tools should be basically the same.
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!!!!
Thanks for sharing this tutorial.
thanks for your video, It was useful and easy to follow
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
Great contribution :)
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!
Thank you so much. It helped a lot. God bless you.
I can't add a unique ID field. The window does not show my unique ID field for my OLS.
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.
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.
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.
Thank you very much for that amazing video lecture.
Thank you for sharing your knockledge with us. Dropbox links doesn´t work. Could you replace it to help us to download dataset, please?
www.dropbox.com/s/8c2jfcs8d82emoc/Lab_GWR.zip?dl=0
thanks for sharing！it's really help！
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?
Excellent
Well Done
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
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.
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..
Can you please share the step by step of how to prepare the data from Hot Spot analysis to GWR dataset?
ndprasad530@gmail.com
I put the URL in the description.
Thank you !!
Very much appreciated!
Excuse me...
if I have point data instead of polygon in micro-area about 9km square. Is it possible to do like this ?
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?
Great video. Could you please share dataset with us?
www.dropbox.com/s/8c2jfcs8d82emoc/Lab_GWR.zip?dl=0
Thanks very much
Amazing tutorial! Could I get the dataset for this tutorial? The dropbox link is not found, thank you.
www.dropbox.com/s/lnn1pvnagqyyd3z/Lab_GWR.zip
Thanks I updated the link in the description as well
Very nice, thank you very much. Congratulations, very good tutorial!!
Moulay Anwar Sounny-Slitine www.dropbox.com/s/nzwguxllynomc8n/Lab_GWR.zip?dl=0
sir, please send the pdf of this tutorial. I am not able to download it. plz send in this email id.
You are welcome.
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 UA-cam. In any case, thanks for sharing!
Good teacher
Thanks!
Just want to ask if it still works on ArcGIS 9.3? tnx
thanks for your video, can you please share data,I want to test some thing. Thanks
Useful Information
Very interesting
can you please share data used forn this , i want test some thing. Thanks
www.dropbox.com/s/nzwguxllynomc8n/Lab_GWR.zip?dl=0
flaco haceme la otra parte
A HUEVO BROTHER..