Great effort, detailed explanation also a most straightforward way to understand the Morans' I High High and Low Low relation. Thanks for the excel sheet
Thank you for the video; At 13:40 you said that values of each element should be same in order to sum up to 1, but in some rows there are some different numbers like 0,125 and 0,13. waiting for ur explanation, thanks
The video was good and give me further understanding about Moran's I but can you make a video where Spatial Autocorrelation using Distance-Based Weights?
Thank you for the course, it is helping me to understand Spatial econometrics, and I will use the knowledge in my thesis. I am having some difficult to fallow the excel the calculation of the matriz of all the possible combination (Xi-xbar)(xj-xbar).
It's an excellent introduction and it's well presented and an excellent introduction to GeoDa. Congratulations on an excellent training methodology. It's impressive. I recommend watching it for an aspirant spatial analyst, bear in mind there is also HIgher Order Contiguities as well as Distance band, and Spatial Weights as Distance Function, K Nearest Neighbors, etc etc Lastly, perhaps a need to mention the difference between Local and Global Spatial Autocorrelation and Gearys C
Hello, thank you for this great series of video. Could you please explain why the Expected value of Moran's I under no spatial correlation is (-1)/(N-1) ? This is not obvious for me.
Are there any restrictions on using Moran's I. I thought there was a minimum on the number of cells needed? Also when you are going over the Moran's I formula is x-bar the average of the entire dataset or just the associated row for the cell? Thank you!
I have not heard of a minimum number of cells- if you find a source on that, please let me know! Certainly, below some very small number, it would be utterly nonsensical to run... Xbar is the grand mean of all observations- sorry if I wasn't clear enough on that.
I have never heard of a minimum number of cells (i.e., polygons or areas) needed. I suppose the minimum number would be two, though I'm not sure if a two-cell map would tell us much...
No, this is how one variable correlates with itself in neighboring regions (univariate). I am not familiar with bivariate spatial correlation; at that point I'd be looking at spatial regression.
Nothing that I can think of. Normally correlation is between TWO variables. The prefix auto- means "with itself". In time series autocorrelation means the correlation of one variable with itself over time. With spatial, it means correlation of a variable with itself in neighboring areas. Although it would be possible to spatially correlate two variables, I don't think it would make sense to call this a spatial correlation. Make sense?
Yes, please check out my list of videos at spatial.burkeyacademy.com so you don't miss anything! The video you want is here, about 16 minutes in: th-cam.com/video/_bnorgXbSG4/w-d-xo.html
-why is (xi - xbar)(xj - xbar) equal to (xi - xbar) square? wasn't that the j is neighbor with a different value?- Nice video btw, the excel helped!! updated: my own fault!! his video is terrific!!
@@BurkeyAcademy-not in the video but in the excel. Between the (xi - xbar)(xj - xbar) and (xi - xbar) square matrix, theres a box saying matrix below and above are identical.-
Thank you for letting me know! Google has been messing up my Google Drive links! I fixed the link in the video, but here it is (please let me know if you have any other problems!): drive.google.com/file/d/0B3-F8BTZSbH9NEVJM2c1Tkg2SjQ/view
This is a master peace of explaining something. You should be proud of it Burkey!
Thanks! I do like this one.
That was so good . I was tired of searching on the internet. thank you for this masterpiece.
Great video! I learned more with this 25 minute video than with some papers out there.
Glad you liked it!
Mr. Burkey Academy, you are a lifesaver 🙏🙏🙏
Great effort, detailed explanation also a most straightforward way to understand the Morans' I High High and Low Low relation. Thanks for the excel sheet
Huge thanks! Really, this is what I was looking for. I really needed to see an actual example.
Wow! Ive been struggling to understand Row Standardization this whole semester. After watching this video it finally clicked!
Very cool! Glad you get it!
It becomes easy to understand with ur explanation, need more videos on spatial stat 😍
You are a great teacher! Thank you for sharing your knowledge, this was very useful to me.
Thank you for the video;
At 13:40 you said that values of each element should be same in order to sum up to 1, but in some rows there are some different numbers like 0,125 and 0,13. waiting for ur explanation, thanks
The only difference might be the width of the column rounding .125 to .13
Thank you very much for a very thorough explanation. It truly reinforces the understanding.
This is so awesome. I can't thank you enough
It was awesome bro... clarified a lots of confusion that I had! :)
Excellent - A must watch
The video was good and give me further understanding about Moran's I but can you make a video where Spatial Autocorrelation using Distance-Based Weights?
Thank you for the course, it is helping me to understand Spatial econometrics, and I will use the knowledge in my thesis. I am having some difficult to fallow the excel the calculation of the matriz of all the possible combination (Xi-xbar)(xj-xbar).
Give a shout if you still need assistance
It's an excellent introduction and it's well presented and an excellent introduction to GeoDa.
Congratulations on an excellent training methodology. It's impressive.
I recommend watching it for an aspirant spatial analyst, bear in mind there is also HIgher Order Contiguities as well as Distance band, and Spatial Weights as Distance Function, K Nearest Neighbors, etc etc
Lastly, perhaps a need to mention the difference between Local and Global Spatial Autocorrelation and Gearys C
Great way of explaining it!
Hello, thank you for this great series of video.
Could you please explain why the Expected value of Moran's I under no spatial correlation is (-1)/(N-1) ? This is not obvious for me.
Thanks for this video!
讲得很清楚，谢谢
Are there any restrictions on using Moran's I. I thought there was a minimum on the number of cells needed? Also when you are going over the Moran's I formula is x-bar the average of the entire dataset or just the associated row for the cell? Thank you!
I have not heard of a minimum number of cells- if you find a source on that, please let me know! Certainly, below some very small number, it would be utterly nonsensical to run... Xbar is the grand mean of all observations- sorry if I wasn't clear enough on that.
I have never heard of a minimum number of cells (i.e., polygons or areas) needed. I suppose the minimum number would be two, though I'm not sure if a two-cell map would tell us much...
Thank you, professor, for this wonderful explanation
Is this what is called bivariate moran's ?
No, this is how one variable correlates with itself in neighboring regions (univariate). I am not familiar with bivariate spatial correlation; at that point I'd be looking at spatial regression.
Many thanks Professor@@BurkeyAcademy
Muito obrigado!
Thank you
Thank you so much
what is the difference between the spatial autocorrelation and spatial correlation?
Nothing that I can think of. Normally correlation is between TWO variables. The prefix auto- means "with itself". In time series autocorrelation means the correlation of one variable with itself over time. With spatial, it means correlation of a variable with itself in neighboring areas. Although it would be possible to spatially correlate two variables, I don't think it would make sense to call this a spatial correlation. Make sense?
Can you please calculate the spatial correlation in R?
Yes, please check out my list of videos at spatial.burkeyacademy.com so you don't miss anything! The video you want is here, about 16 minutes in: th-cam.com/video/_bnorgXbSG4/w-d-xo.html
-why is (xi - xbar)(xj - xbar) equal to (xi - xbar) square? wasn't that the j is neighbor with a different value?-
Nice video btw, the excel helped!!
updated: my own fault!! his video is terrific!!
Can you tell me at what time I said that these are equal? I shouldn't have, and can't find where I did. Thanks!
@@BurkeyAcademy-not in the video but in the excel. Between the (xi - xbar)(xj - xbar) and (xi - xbar) square matrix, theres a box saying matrix below and above are identical.-
NVM i got it, just realised how to read the formular
OK, glad you liked it! If there is anything I can make clearer, please let me know. Cheers!
The file doesn't exist anymore. Could you please upload a new one?
No worries, just found it on your web. Many thanks!
Thank you for letting me know! Google has been messing up my Google Drive links! I fixed the link in the video, but here it is (please let me know if you have any other problems!): drive.google.com/file/d/0B3-F8BTZSbH9NEVJM2c1Tkg2SjQ/view