Hello Diogo, looks like it’s been 5 years since you posted this and it’s still helping people! I would never had figured out some of the idiosyncrasies to formatting without your help! Thanks so much!
wow... thank u for your channel. I was looking this information months ago but I found nothing, and thanks for share your knowledge in here which is very important to me. Thanks.
Thank you for providing the spatial panel analysis tutorial. But how to display the residuals from the model so that classical regression assumptions can be tested? And how we get LM test and LR test?
Hi Diogo, thank you very much for this excellent tutorial. I followed your instructions carefully, but when I got to the part where I issue the command "spmat dta W m*, ...", Stata gives me the following error message: "main diagonal must contain zero elements". I'm not sure why I'm getting this message because I downloaded your weights file and compared it to mine and they look similar...
3 роки тому
Hi. The matrix must have only 0s in it's diagonal, otherwise the spatial units are neighbours of themselves, which doesn't make sense. For example, the matrix must have its first value as a 0, because the first observation/spatial unit can't be a neighbour of itself. Imagine a matrix of Portugal and Spain (they are neighbours). The matrix would be: 0 1 1 0
hi, thanks for the video, could you tell me how to download the matrix of weights from excel in the stata (in excel, the matrix of trade flows by region)
Is there away someone can reach out to you, thank you. This video is really great but could use more guidance as I am a beginner on this and no one in my circles has an idea
Hi Diogo, Thanks for uploading this video. You are awesome. However, I have a question. Is it possible to get direct, indirect, and total effects for spatial panel model? Such as Spatial panel durbin model or spatial panel lag dependent variable? Your answer will be appreciated. Thanks
The paper he refers to by Belotti & Hughes has the command 'effects' within the main command 'xsmle' to compute these effects. Do refer the paper, it is very detailed!. Thanks @Diogo for this amazing tutorial.
Hi Diogo. Thanks for the excellent tutorial video, it describes the whole process of learning spatial regression and it helped me a lot. When I convert shp > gal > txt files, I have a question. There are some islands in my research data, how can I solve the problem that the islands have no neighbors? Thank you!
Greetings Diogo, Thank you for uploading this; it is really helpful. I have a question. What if we are not considering every district or municipality? What if we have data for only 30 out of 35 areas? How should we go about that?
8 місяців тому
Hey. If you have missing/unbalanced data, check the paper I reference at the end of the video (the link is in the description). They explain your options.
I've excluded five states from my analysis because I lack any information about them. The paper suggests imputation, but unfortunately, that's not feasible in my case. Among the excluded areas, two are islands with no shared boundaries with any other states, two share a border with just one state each, and one shares borders with three states. I'm wondering if there's an alternative approach to address this issue. Specifically, I'm considering dropping these five states at the time of creating the spatial weight matrix. Would this be a viable solution?
8 місяців тому
@@nikeesilayach3685 It may be a viable solution to exclude States from your study. If you are going for a contiguity spatial weights matrix (as I show in this video), by definition, you must exclude islands. Regarding the 3 states for which you have missing data, assuming you have no way of getting hold of the missing data and you cannot find a way to input the missing data, in my opinion, you are left with no good options. You can exclude those 3 states from your analysis but this can potentially limit the explanatory power / scope of your study (because you are in effect pottentially ignoring significant spatial interactions between states).
Very helpful. I just wanted to know, If I have any neighborless spatial unit, then what I should do? I don't want to remove them from the analysis because they are important. TIA
Hello Diogo, I find your video to be very helpful. I performing a similar analysis with Chicago census tract data and I keep getting the error that " main diagonal must contain zero elements". I am finding it difficult to identify where the issue is in the matrix. Is there a way to directly contact you for assistance? Thank you.
Рік тому
Hello, The matrix must have only 0s in its diagonal, otherwise the spatial units are neighbours of themselves, which doesn't make sense. For example, the matrix must have its first value as a 0, because the first observation/spatial unit can't be a neighbour of itself. Imagine a matrix of Portugal and Spain (they are neighbours). The matrix would be: 0 1 1 0 Check your matrix carefully!
Hello Diogo Tome, I tried replicating the do-file you wrote, but it does not work for the data I am using. I use provincial-level data in Indonesia. Can you provide input? I would be delighted if you could help. I am very willing to provide the data da do-file that I have written. Thank you.
Hi Diogo, your video is helpful. Thanks. I use xsmle, but how to do LM error and LM lag tests for each effect? fixed effect and time fixed effect for example
hi diogo, thankyou for this really informative video. I am actually encountering a problem i.e. the univariate local moran 1 is not showing the island data but the .gal file is showing it and even after i edited it and added artificial neighbours to the islands, stata is showing invalid data
Dear Diogo Tomé, thanks for the very usefull video. I have a doubt and I would like to address. Do STATA compute the Global and Local Durbin Models. The local is known as Spatial Durbin Error Model. Thanks.
In the following command "xsmle Y X1 X2 X3, wmat(W) model(sdm) fe type(ind) nolog" I assume this is spatial fixed effect model. inside the parenthesis, what does "ind" stand for? For SDM direct, indirect, total effects model, how can we define short-run and long-run? I would appreciate if you refer any material.
thanks for useful video.I have question about regression with two matrix, I mean one regression with two lag matrix and than two rho coefficient, plz help me how could run and find codes for stata, thank you very much.
Hi Diogo, Thank you for the video. Do you have any knowledge on the spm package? It extends the xsmle that is able to estimate spatial models using two spatial weights. 'help spm' in stata command.
For SDM with spatial fixed effects [data transformed according to Lee and Yu (2010)], there is a error report "variable wmat not found". For all Hausman Test, error report "_xsmle_hausman_ml(): 3200 conformability error" and ": - function returned error" shows. How may I fix these report? Thanks in advance.
Hi Diogo, Thanks for your video! it's very helpful! I have a question, is the "id" in the weights="mun" file in the data panel? I mean, should we have the same code in the both files?
2 роки тому
Correct. The IDs have to represent the same thing both in weights file and in the excel data
This video is really helpful. But when I tried to import a 283*283 contiguity matrix from Geoda into Stata, Stata showed that my system has run out of application memory. Do you know how to fix it? Thank you.
Hai Diogo, Thanks so much for your presentatiton and sharing this video. It's amazing. By the way, do you have tutorial about estimate spatial dependence destination-based model, for flow data like migration? can u share please...
Diogo, bom dia. Tenho duas dúvidas que preciso tirar com você, pode me ajudar? estou acompanhando sua análise, porém fiquei perdido em entender, pois primeiro você renomeou as variáveis para m (até ai eu entendi), mas depois quando montou o painel, utilizou as variáveis v (quais seriam essas variáveis, não consegui entender de onde tirou essas variáveis, poderia esclarecer por favor?), obrigado, ótimo vídeo
3 роки тому
Olá :) mudei de v para m o nome das colunas da matriz espacial para que mais tarde, quando usei as variáveis para estimar os modelos espaciais, não houvesse conflito. Só tive de fazer isso porque escolhi chamar às variáveis v1, v2 e v3 (vê o ecrã no minuto 19:40)... mas se tivesse dado outro nome às variáveis, desde que depois fizessem match com o código (vê o ecra no minuto 35:20), não ia haver problema. Para estimar qualquer modelo econométrico, precisas de dados de variáveis (no exemplo sao as variáveis v1 v2 e v3). Adicionalmente, por aqui estarmos a estimar modelos espaciais, precisas também da matriz espacial (em que às colunas dei o nome m) Qualquer dúvida diz
Thank you for sharing your knowledge. I would like to ask you a question if you don't mind. I tried to execute the spwmatrix command in order to export the weight file to txt. But, it didn't work. A message "J(): 3900 unable to allocate real [5300108,5300108] -- spwmatrix_CalcSPweightM(): - function returned error -- : - function returned error" just showed up... I have 5565 Brazilian municipalities in the sample. Can you help?
I figured out the problem in my case. My ID variable started with large numbers, like 1100015. So, I changed the sequence, starting by 1, 2,..., 5565. It worked!
When I ran the code, the following error happend with my own data " _xsmle_hausman_ml(): 3200 conformability error : - function returned error" . Do you know how can I peform the Hausman test in this case?
Hello. Thank you for the very informative video. I'm trying to use the codes to model viral circulation. I encountered an error stating "matrix not square". I think I followed your instructions on how to import the weights in stata. Can you perhaps suggest how to overcome this error? Thank you very much.
6 років тому
Michael Bacus hey thanks! Make sure the data you're using for the weights has no problems... If you're using a gal file for your weights, I'd take a look into that using notepad. Obviously matrixes need in a rectangular or square form...
Thank you. I've fixed my data already. Can I get your email if it's alright? I have a few more questions regarding spatial models. I'm still a newbie in modelling and I'd appreciate it if you can advise me on certain things. Thank you very much once again.
It will be great if you did a video for Spatial unbalanced panel
7 років тому
I appreciate your suggestion. I should have said in the video that you cannot run the xsmle command with unbalanced panels. If you do, you get this message: Error - the panel data must be strongly balanced with no missing values. So, if you have missing values in your data (you can see if you do by typing "mdesc" in Stata), you have to deal with them. You can use the multiple imputation method (in Stata: mi impute regress) to fill in the missing values with "plausible" values. Here's a video showing how to do just that: ua-cam.com/video/i6SOlq0mjuc/v-deo.html Also check this link explaining how to run the xsmle command while using an mi estimate: www.econometrics.it/?p=383 You can also see the paper I linked in the description of the video, it adresses unbalanced panels with an example.
Thanks a lot, sir! Are there any subtleties when using matrix based on distances? Will we see extention of this video with weights matrix based on distances?
6 років тому+1
Hey thanks for your comment. About your first question. When you create a spatial weights matrix in Geoda based on contiguity, the matrix is saved as a .gal file. When you create one based on distance, it saves it as a .gwt file. That's the first subtlety. If you scroll down and read the other comments you will find one telling you how to import a .gwt file into Stata. If there are other subtleties, it will be on how to actually use the distance weights in Stata, after you have imported it. At that point if you struggle I say just search on google for the answers cos I honestly don't have them at the top of my head. Second question. Unlikely, maybe you can do it? :P
it's a great video bro !! however, is the spwmatrix command only works at Stata 12 above? I was following your guidance, yet i could not execute the spwmatrix import...sorry for my very basic question, I am a newbie here in stata...
Hi! When I am importing the gal file the system saya : 3499 spwmatrix_CalcSPweightM() not found I have installed the spwmatrix but it still doesn't work.
5 років тому
sorry man I don't know how to trouble shoot that...
Hi! Thank you for your information! I used your methods and it kind of works, but the problem is my gal file contains 2235*2235 obs, which fail to generate a txt. file by running the spwmatrix and xport command. In this case what should I do? I appreciate your help!
PS: by running the command of spwmatrix in STATA, I got "too many variables specified error in line 2,235 of file r(103);"
5 років тому
@@raymondhuang4741 Hey, your problem may have to do with limitations on the Stata you're using. You need to be running on Stata/MP for it to accept that matrix size. Quoting from the 2017 paper I link in the description (page 7): "It is worth noting that the maximum dimension of a single Stata matrix depends on Stata’s flavor: 40 × 40 (Small) 800 × 800 and 11,000 × 11,000 (SE or MP). In order to overcome this limitation, bigger matrices have to be specified as spmat objects."
Hello Sir, Can you guide how can i get weights for countries? i want to do spatial analysis for my thesis. total 40 countries and 4 regions. How can i make or get shape and weights?
thanks for useful video. I want to estimate spatial panel var model, but I can't find it's command. Can you help me?
7 років тому
Hey, thanks for your comment. You can use the pvar package to estimate panel VAR models in Stata. As for spatial panel VAR models, I'm afraid there isn't a simple command that will do the trick. Check out these 2 articles, maybe they can be of use to you: - "Ageing shocks and short-run regional labour market dynamics in a spatial panel VAR" - "Ripple effects of house prices: considering spatial correlations in geography and demography" Good luck!
Caro Diogo, obrigado pelo vídeo. De grande valor para todos! Estou com um problema em relação à importação da matriz do GeoDa. Tentei esse código: spwmatrix import using "$data\Weight.gal", wname(Wcontig) xport(Wcontig, txt), mas tem dado erro (aparece a mensagem file Wcontig.txt could not be opened). Você tem alguma ideia do que possa estar acontecendo?
Hi sir, i Am a first year PhD sudent. I want to test the spatial effects of entrepreneurship on economic development for the middle East and North Africa region. Is that possible? If yes, I can't extract the shapefile for the whole region! Please I need your help
5 років тому+1
It is possible. First you have to define how you're going to measure "entrepreneurship", because that's going to be your dependent variable (and you probably will want to include "entrepreneurship in neighbouring spacial units" as an independent variable too). Your spatial units could be countries or --probably better-- cities. A distance matrix will work best in this case. It all depends on how you frame your goal. Good luck!
@ thank you sir. If you Are interested we can publish this article together. Just send me your e-mail. Otherwise I am very grateful for your help. Thank you again.
Hey Diogo. Great video I have successfully created the weight matrix for my country. Can even run the basic reg. command. However, whenever I try to use one of the spatial models (example), the following error occurs. Example: xsmle dun popdens month_6, wmat(W) model(sar) fe nolog
My Panel is Strongly ballanced. *: 3200 conformability error _xsmle_est(): - function returned error : - function returned error My Stata is fully up to date and I literally have tried every single link on google. Would be thankful for any sort of advice as to what might be causing this issue. Lukas
Hi: I have a problem again. I am very satisfied if you can help me. The weight matrix I use is the threshold matrix. I did not have any problem for the lower values, but when I made it for 400km it gave me an error.(Initial values not feasible R (1400))...I how to this problem..
7 років тому+1
Hi. I suggest you read what's been said in these 2 links: 1. www.statalist.org/forums/forum/general-stata-discussion/general/1400685-problem-with-weights-matrix-in-spatial-panel-model-xsmle 2. www.statalist.org/forums/forum/general-stata-discussion/general/1398048-xsmle-initial-values-not-feasible-numerical-overflow
Hi.. I want to guess with the gmm. Could you help in estimating the gmm for the spatial panel model in the stata program.
7 років тому
Hey Kübra, sorry for the late reply. The xsmle command uses the Quasi-Maximum Likelihood (QML) estimation method. Unfortunately, I'm not aware of any way to estimate spatial panel data models in Stata using Generalized Method of Moments (GMM) techniques... But using the QML method shouldn't be a problem: Soro, W. L.; Zhou, Y. and Wayoro, D. in “Crash Rates Analysis in China using a Spatial Panel Model” (2016, p.2) say that it's recognized that “the ML and QML methods are so far the most appropriate methods to fit spatial panels”.
www.statalist.org/forums/forum/general-stata-discussion/general/1349975-spregxt-new-stata-module-econometric-toolkit-to-estimate-spatial-panel-regression-models (SPGMM) Spatial Panel Autoregressive Generalized Method of Moments Model
Hello Diogo, thank you so much for this informative tutorial! I have a question. Your data is perfectly balanced. Can your code/weights be applied to an unbalanced data set? Thank you again.
4 роки тому+1
Xsmle won't run if you have unbalanced panel data. If you search the comment section you will find some people raising this issue and hopefully some helpful replies. There are work arounds. You can transform the missing data using mi. The xsmle paper I link in the description has a section showing how to do it. That said, if I were you I'd make sure that data I was using was balanced.
so I get "initial values not feasible" in my SAR models...what does this mean and how can I fix it?
7 років тому
I suggest you do a google search like this: "Initial values not feasible" xsmle (with the " " so it searches exactly those terms). You can usually find the solution to any problem in forums. Check these 2 links: www.statalist.org/forums/forum/general-stata-discussion/general/1400685-problem-with-weights-matrix-in-spatial-panel-model-xsmle www.statalist.org/forums/forum/general-stata-discussion/general/1398048-xsmle-initial-values-not-feasible-numerical-overflow Good luck!
Hi bro thanks for your useful video. But i have a problem when I ran mine. I ran step by step according to your commands with my data, my when I ran 'estimating fixed-effects model to perform Hausman test', it reminded me " _xsmle_hausman_ml(): 3200 conformability error". However, I used the same weight matrix for SDM and they were all okay. Can you please tell me what happened? Thanks
6 років тому
I did some google searching and your conformability error is probably because of the size of your spatial weights matrix. Maybe you can ask Federico Belotti (he wrote the xsmle command) to help you: federico.belotti@uniroma2.it
Hi... could you help me with unbalanced data please? I tried to follow the instruction using Beloti and friends paper but I stacked because when I enter this syntax, mi impute regress x1 = z, add(50) rseed(12345), it said that variable z not found. Could you please to help me with this? Thank you.
7 років тому
Hi. First of all I would advise you to stay away from unbalanced panel data sets. You may be better off removing the variables which have missing values altogether. Having said that, you are stuck because you, unlike the authors in the example which you are following, did not define the variable "z". You don't have to define a "z" variable to proceed, though. You should use the variable that makes sense for your data. Type in Stata: help mi impute regress Also, check this video: ua-cam.com/video/i6SOlq0mjuc/v-deo.html I hope it helps!
ohh ok Nuno... the problem is most of main control variables have missing values, so it I have to deal with it. Btw, thank you for the link.
7 років тому
Maybe there are alternative control variables which can serve you just as well :p Notice that the multiple imputation method (mi impute regress) deals with unbalanced panels by filling in missing values with "plausible" values. That is to say, using the mi strategy you will end up with balanced data. You can only run the xsmle command with balanced data. It may help knowing that, when you are using a .dta file in Stata and you run the command "mdesc", Stata will present a table with the number of missing values for each variable in that .dta file. Thanks for your comment!
Hi, thanks for the video! I'm experiencing the error invalid 'GeoData' r(198); when I try to import the .gal file using this syntax. spwmatrix import using "/Users/cameronmartin/Desktop/UK GeoData/GeoDa/Test.gal", wname (SPW) xport(SPW, txt) Thanks
6 років тому
Did you find a way to make Stata find your .gal file? That's a basic syntax error. Sorry I took so long to reply friend. I didn't notice your comment til now
Hi, Diego thanks for your reply. No, I never found out how to get around this error on STATA. I ended up using a smaller dataset on R. Is it possible for you to help me work through this on STATA? Thanks
6 років тому
Try moving your Test.gal file to the desktop and run this command in Stata: spwmatrix import using C:\Users\cameronmartin\Desktop\Test.gal, wname(SPW) xport(SPW, txt)
please I need help ! for my work I have only one regressor so when I execute the code I obtain ;"to few variables ".. should I HAve 2 variables at least ?
7 років тому
To run a regression you need 1 dependent variable (in the video, v1) and 1 or more independent variables / regressors (in the video, v2 and v3). So, you only need 1 regressor, but you should have at least 2 variables (1 dependent variable and 1 regressor). Did you include the dependent variable? The purpose of any regression is to understand how much the dependent variable can be explained by the independent variable(s). It's an attempt to predict a value of your dependent variable based on values of your independent variable(s). Spatial regressions (xsmle) serve the same purpose, only you're also accounting for spatial dependency (based on the spatial weighs matrix). Notice people can refer to the independent variable as the eplanatory variable / regressor / predictor variable / controlled variable / manipulated variable / input variable. And to the dependent variable as the explained variable / response variable / predicted variable / measured variable / experimental variable / outcome variable / output variable.
I dont know why my stata say"main diagonal must contain zero elements"when i print "spmat dta W m*,normalize (row)"
5 років тому
The diagonal elements of the weights matrix have to be equal to 0 to exclude self neighbors! If there are "1"s in your main diagonal, there's something wrong with the matrix.
Hİ! I want to apply lagrange multiplier test and robust lagrange multiplier test for spatial panel data analysis. Can ı do it with stata. Can you help me find CODs.I am very satisfied if you can help me.
7 років тому
You can't do the LM tests using xsmle. I don't think there is a command in Stata that allows you to do those tests for spatial panel models. You can do them in R though. If you are interested, google "splm: Spatial Panel Data Models in R"
helloo. i want to estimate general nesting model (elhorst). can you help me to estimate gns model?
7 років тому
Hi. I don't know that specific literature so I can't really help you. Having said that, at a quick glance, as far as I can tell, there isn't a single "general nesting model". Some of the "nested models" used are models that can be estimated with the xsmle command in Stata (SARAR and SDM). Check this link: old.aecr.org/images/ImatgesArticles/2012/3/06_PETER.pdf
ıf my files .gwt how to read it to stata. .ı used '' spwmatrix import using d:/non.gwt,wname(non) xport(non,txt)'' but spwmatrix_CalcSPweightM(): 3200 conformability error
7 років тому
The spwmatrix command does not allow you to import spatial weights from a .gtw file... You can do it using the spmat command. Follow the instructions in this link: www.stata.com/statalist/archive/2011-10/msg01385.html
my file name is 'non'. and it is save D disk. ı used 'spmat import ' but STATA give error 'cannot open non.GWT'. Can you help me this subject? I think I wrote wrong.Can you write this command as an example.
7 років тому
Notice what I did in 11:40. Try changing your working directory to Desktop before running the command. Let me know if it fixed your problem.
7 років тому
After you change your Working Directory to desktop, you have to let Stata know where your non.gwt file is. If your non.gwt file is in the Desktop, you run this command: spmat import non using C:\Users\*YourUserName*\Desktop on.gwt, geoda
yes,,i have,,,when i dowloded the data from the site i get 4 shape_file,with 1,20,306,and 4260 municipalities,,,,my question was how did you get the shpa_file with 287 municipalities. thank you.
6 років тому
Just watch what I do on second 2:55 of this video. What I uploaded to google drive is exactly what I used in the video... So the municipalities.shp file only has 278 municipalities
Nuno Álvares Pereira, this video is very good! I wanted to take this opportunity to ask for your help. I have to deliver an article on the 06th, but I can not generate my weight matrix. I tried to follow your do-file, but I could not run the first command. Help me please! The error that appears is: spwmatrix import using C:\Users\Jordana\Desktop\Stata\BRMUE250GC_SIR.gal, wname(weights) xport(weights, txt) J(): 3900 unable to allocate real [5300108,5300108] spwmatrix_CalcSPweightM(): - function returned error : - function returned error But what is the necessary memory? My computer has 1T of storage in HD, 8G RAM, i3 processor.
![Lp. Сердце Вселенной #60 РОЖДЕНИЕ ЛОЛОЛОШКИ [Финал] • Майнкрафт](http://i.ytimg.com/vi/YoR0pAV9FVQ/mqdefault.jpg)
Hello Diogo, looks like it’s been 5 years since you posted this and it’s still helping people! I would never had figured out some of the idiosyncrasies to formatting without your help! Thanks so much!
Thank you so much from Turkey for telling us what we couldn't find in the books.
I don’t know what my research would be like without your guidance here. Thank you from Indonesia.
Bahas apa mbak? Saya juga masih belajar spasial di stata
Hi Diogo. With your guidance with this tutorial , I managed to publish my latest article on a very important journal in the field.
Thanks a lot.
Thanks for sharing and congrats! I'm happy for you.
The most useful video on UA-cam I've ever watched. Thank you so much!
wow... thank u for your channel. I was looking this information months ago but I found nothing, and thanks for share your knowledge in here which is very important to me. Thanks.
Thank you so much from Cambridge!! You are literaturally a hero !
fantastic! i really appreciate your generosity and knowledge you have shared with us. Thank you a lot.
such a great tutorial, thanks helped me a lot for my phd thesis
It is very kind of you to post this helpful video.
Es único este video en cuanto al tema, explicación y tratamiento de la información. Muchas gracias Nuno por subirlo.
I think you helped me save my dissertation. Very informative! Thank you
thank you very much for this highly informative video. It helped me a lot in running spatial panel data models in stata.
Thanks, sir, for such an informative video
Thank you so much for this fantastic teaching! I have learned so many useful things.
Thanks Lisa
Terima kasih banyak Sir. Sangat membantu sekali dalam penelitian saya
Wow, it was a perfect example!! Thank you very much!!
Thank you very much :)
thank you for all of data files.
Thank you for providing the spatial panel analysis tutorial. But how to display the residuals from the model so that classical regression assumptions can be tested?
And how we get LM test and LR test?
This was incredibly helpful. Thank you!!!!!
You are great!! Your lecture is fantastic!
You are great too :p thanks!
Obrigada, Diogo! Muito bom o seu vídeo, o melhor que já vi.
Muito obrigado senhora
Thank you soo much. It is very helpful. please do similar video for cross section data and if we can use it for spatial regression.
Hi. Thank you. For cross section data you don't need to use stata. You can run the regression in Geoda. Good luck!
thanks for your sharing i learnt a lot
Hi Diogo, thank you very much for this excellent tutorial. I followed your instructions carefully, but when I got to the part where I issue the command "spmat dta W m*, ...", Stata gives me the following error message: "main diagonal must contain zero elements". I'm not sure why I'm getting this message because I downloaded your weights file and compared it to mine and they look similar...
Hi. The matrix must have only 0s in it's diagonal, otherwise the spatial units are neighbours of themselves, which doesn't make sense.
For example, the matrix must have its first value as a 0, because the first observation/spatial unit can't be a neighbour of itself.
Imagine a matrix of Portugal and Spain (they are neighbours). The matrix would be:
0 1
1 0
Thank you for the informative video!!
Thank you sir
hi, thanks for the video, could you tell me how to download the matrix of weights from excel in the stata (in excel, the matrix of trade flows by region)
Is there away someone can reach out to you, thank you. This video is really great but could use more guidance as I am a beginner on this and no one in my circles has an idea
Hi Diogo, Thanks for uploading this video. You are awesome. However, I have a question. Is it possible to get direct, indirect, and total effects for spatial panel model? Such as Spatial panel durbin model or spatial panel lag dependent variable? Your answer will be appreciated. Thanks
The paper he refers to by Belotti & Hughes has the command 'effects' within the main command 'xsmle' to compute these effects. Do refer the paper, it is very detailed!. Thanks @Diogo for this amazing tutorial.
It's a great video!, thanks
Hi Diogo.
Thanks for the excellent tutorial video, it describes the whole process of learning spatial regression and it helped me a lot. When I convert shp > gal > txt files, I have a question.
There are some islands in my research data, how can I solve the problem that the islands have no neighbors?
Thank you!
Greetings Diogo,
Thank you for uploading this; it is really helpful. I have a question. What if we are not considering every district or municipality? What if we have data for only 30 out of 35 areas? How should we go about that?
Hey. If you have missing/unbalanced data, check the paper I reference at the end of the video (the link is in the description). They explain your options.
I've excluded five states from my analysis because I lack any information about them. The paper suggests imputation, but unfortunately, that's not feasible in my case. Among the excluded areas, two are islands with no shared boundaries with any other states, two share a border with just one state each, and one shares borders with three states. I'm wondering if there's an alternative approach to address this issue. Specifically, I'm considering dropping these five states at the time of creating the spatial weight matrix. Would this be a viable solution?
@@nikeesilayach3685 It may be a viable solution to exclude States from your study.
If you are going for a contiguity spatial weights matrix (as I show in this video), by definition, you must exclude islands.
Regarding the 3 states for which you have missing data, assuming you have no way of getting hold of the missing data and you cannot find a way to input the missing data, in my opinion, you are left with no good options.
You can exclude those 3 states from your analysis but this can potentially limit the explanatory power / scope of your study (because you are in effect pottentially ignoring significant spatial interactions between states).
Very helpful. I just wanted to know, If I have any neighborless spatial unit, then what I should do? I don't want to remove them from the analysis because they are important. TIA
Thank you very much
I have to estimate SDM ans SEM model without log, it's possible to estimate that??
Hello Diogo, I find your video to be very helpful. I performing a similar analysis with Chicago census tract data and I keep getting the error that "
main diagonal must contain zero elements". I am finding it difficult to identify where the issue is in the matrix. Is there a way to directly contact you for assistance? Thank you.
Hello,
The matrix must have only 0s in its diagonal, otherwise the spatial units are neighbours of themselves, which doesn't make sense.
For example, the matrix must have its first value as a 0, because the first observation/spatial unit can't be a neighbour of itself.
Imagine a matrix of Portugal and Spain (they are neighbours). The matrix would be:
0 1
1 0
Check your matrix carefully!
@ thanks, I will go through it again. Thank you.
Hello Diogo Tome,
I tried replicating the do-file you wrote, but it does not work for the data I am using.
I use provincial-level data in Indonesia. Can you provide input? I would be delighted if you could help. I am very willing to provide the data da do-file that I have written.
Thank you.
Hi Diogo, your video is helpful. Thanks. I use xsmle, but how to do LM error and LM lag tests for each effect? fixed effect and time fixed effect for example
hello there, All the examples, beforing, are to balanced panel, so you could show us as estimating an unbalanced panel data in Stata?
hi diogo, thankyou for this really informative video. I am actually encountering a problem i.e. the univariate local moran 1 is not showing the island data but the .gal file is showing it and even after i edited it and added artificial neighbours to the islands, stata is showing invalid data
Dear Diogo Tomé, thanks for the very usefull video. I have a doubt and I would like to address. Do STATA compute the Global and Local Durbin Models. The local is known as Spatial Durbin Error Model. Thanks.
So cool!! Thanks very much~
Thank you!
Hi, When I plug in my data I am getting 3200 conformability error
very well done, thank you so much!
Thank you!
so i'm pretty new to Stata - is it possible to create a variable list and then run the models on a list instead of copy pasting the variable list?
Aurélie Shapiro yes just work on a .do file and then run it!
Without a shape file what is the procedure to create a weight matrix only by using panel data
Please give the code. it will very helpful to me
In the following command "xsmle Y X1 X2 X3, wmat(W) model(sdm) fe type(ind) nolog" I assume this is spatial fixed effect model. inside the parenthesis, what does "ind" stand for? For SDM direct, indirect, total effects model, how can we define short-run and long-run? I would appreciate if you refer any material.
thanks for useful video.I have question about regression with two matrix, I mean one regression with two lag matrix and than two rho coefficient, plz help me how could run and find codes for stata, thank you very much.
What should I do if the gspre model has problems with initial values?
Hi Diogo, Thank you for the video. Do you have any knowledge on the spm package? It extends the xsmle that is able to estimate spatial models using two spatial weights. 'help spm' in stata command.
For SDM with spatial fixed effects [data transformed according to Lee and Yu (2010)], there is a error report "variable wmat not found". For all Hausman Test, error report "_xsmle_hausman_ml(): 3200 conformability error" and ": - function returned error" shows. How may I fix these report? Thanks in advance.
Is there any issue with weight matrix?, though I created weight for 48 continental states plus Washington D.C. (USA).
I could solve the weight matrix, but still suffering with Hausman Test. Do you have any advice please?
Hi Diogo,
Thanks for your video! it's very helpful!
I have a question, is the "id" in the weights="mun" file in the data panel? I mean, should we have the same code in the both files?
Correct. The IDs have to represent the same thing both in weights file and in the excel data
@ thank you for your reply
This video is really helpful. But when I tried to import a 283*283 contiguity matrix from Geoda into Stata, Stata showed that my system has run out of application memory. Do you know how to fix it? Thank you.
Thank you soo much, Diogo Tome, It is very helpful, please Do you have any idea of how to convert GeoDa.kwt ( kernel density file) to Stata.dta?
Hai Diogo, Thanks so much for your presentatiton and sharing this video. It's amazing. By the way, do you have tutorial about
estimate spatial dependence destination-based model, for flow data like migration? can u share please...
Diogo, bom dia. Tenho duas dúvidas que preciso tirar com você, pode me ajudar? estou acompanhando sua análise, porém fiquei perdido em entender, pois primeiro você renomeou as variáveis para m (até ai eu entendi), mas depois quando montou o painel, utilizou as variáveis v (quais seriam essas variáveis, não consegui entender de onde tirou essas variáveis, poderia esclarecer por favor?), obrigado, ótimo vídeo
Olá :) mudei de v para m o nome das colunas da matriz espacial para que mais tarde, quando usei as variáveis para estimar os modelos espaciais, não houvesse conflito.
Só tive de fazer isso porque escolhi chamar às variáveis v1, v2 e v3 (vê o ecrã no minuto 19:40)... mas se tivesse dado outro nome às variáveis, desde que depois fizessem match com o código (vê o ecra no minuto 35:20), não ia haver problema.
Para estimar qualquer modelo econométrico, precisas de dados de variáveis (no exemplo sao as variáveis v1 v2 e v3). Adicionalmente, por aqui estarmos a estimar modelos espaciais, precisas também da matriz espacial (em que às colunas dei o nome m)
Qualquer dúvida diz
Thank you for sharing your knowledge. I would like to ask you a question if you don't mind.
I tried to execute the spwmatrix command in order to export the weight file to txt. But, it didn't work. A message "J(): 3900 unable to allocate real [5300108,5300108] -- spwmatrix_CalcSPweightM(): - function returned error -- : - function returned error" just showed up...
I have 5565 Brazilian municipalities in the sample.
Can you help?
I figured out the problem in my case. My ID variable started with large numbers, like 1100015. So, I changed the sequence, starting by 1, 2,..., 5565. It worked!
When I ran the code, the following error happend with my own data " _xsmle_hausman_ml(): 3200 conformability error
: - function returned error" . Do you know how can I peform the Hausman test in this case?
Perfect
Hello. Thank you for the very informative video. I'm trying to use the codes to model viral circulation. I encountered an error stating "matrix not square". I think I followed your instructions on how to import the weights in stata. Can you perhaps suggest how to overcome this error? Thank you very much.
Michael Bacus hey thanks! Make sure the data you're using for the weights has no problems... If you're using a gal file for your weights, I'd take a look into that using notepad. Obviously matrixes need in a rectangular or square form...
Thank you. I've fixed my data already. Can I get your email if it's alright? I have a few more questions regarding spatial models. I'm still a newbie in modelling and I'd appreciate it if you can advise me on certain things. Thank you very much once again.
hai sir, i want to ask, my file weight cant open in stata. this wrong 'file already exist' what should i do sir? pls help me 🥺🥺
Type clear on the command window
So helpful. Thank you
Can anyone please tell me how can I generate a lagged independent variable for spatial panel data in Stata.
It will be great if you did a video for Spatial unbalanced panel
I appreciate your suggestion. I should have said in the video that you cannot run the xsmle command with unbalanced panels.
If you do, you get this message: Error - the panel data must be strongly balanced with no missing values.
So, if you have missing values in your data (you can see if you do by typing "mdesc" in Stata), you have to deal with them.
You can use the multiple imputation method (in Stata: mi impute regress) to fill in the missing values with "plausible" values.
Here's a video showing how to do just that: ua-cam.com/video/i6SOlq0mjuc/v-deo.html
Also check this link explaining how to run the xsmle command while using an mi estimate: www.econometrics.it/?p=383
You can also see the paper I linked in the description of the video, it adresses unbalanced panels with an example.
THANKS
Thanks a lot, sir! Are there any subtleties when using matrix based on distances? Will we see extention of this video with weights matrix based on distances?
Hey thanks for your comment. About your first question. When you create a spatial weights matrix in Geoda based on contiguity, the matrix is saved as a .gal file. When you create one based on distance, it saves it as a .gwt file. That's the first subtlety. If you scroll down and read the other comments you will find one telling you how to import a .gwt file into Stata. If there are other subtleties, it will be on how to actually use the distance weights in Stata, after you have imported it. At that point if you struggle I say just search on google for the answers cos I honestly don't have them at the top of my head. Second question. Unlikely, maybe you can do it? :P
You are an awesome teacher .how to get your "code.do" I am in trouble
here:
drive.google.com/file/d/0By5qMMVfcc4jRHJfNExjUV9pd2c
Dear,
Thank you for your presentation, can we receive stata command by email?
Hi, links are in the video description
it's a great video bro !! however, is the spwmatrix command only works at Stata 12 above? I was following your guidance, yet i could not execute the spwmatrix import...sorry for my very basic question, I am a newbie here in stata...
Hey thanks! Send me the error message you got?
Hi! When I am importing the gal file the system saya
: 3499 spwmatrix_CalcSPweightM() not found
I have installed the spwmatrix but it still doesn't work.
sorry man I don't know how to trouble shoot that...
Hi! Thank you for your information! I used your methods and it kind of works, but the problem is my gal file contains 2235*2235 obs, which fail to generate a txt. file by running the spwmatrix and xport command. In this case what should I do? I appreciate your help!
PS: by running the command of spwmatrix in STATA, I got "too many variables specified
error in line 2,235 of file r(103);"
@@raymondhuang4741
Hey, your problem may have to do with limitations on the Stata you're using. You need to be running on Stata/MP for it to accept that matrix size. Quoting from the 2017 paper I link in the description (page 7):
"It is worth noting that the maximum dimension of a single Stata matrix depends on Stata’s flavor: 40 × 40 (Small) 800 × 800 and 11,000 × 11,000 (SE or MP). In order to overcome this limitation, bigger matrices have to be specified as spmat objects."
Hello Sir, Can you guide how can i get weights for countries? i want to do spatial analysis for my thesis. total 40 countries and 4 regions. How can i make or get shape and weights?
Try here: www.diva-gis.org/gdata
thanks for useful video. I want to estimate spatial panel var model, but I can't find it's command. Can you help me?
Hey, thanks for your comment.
You can use the pvar package to estimate panel VAR models in Stata. As for spatial panel VAR models, I'm afraid there isn't a simple command that will do the trick.
Check out these 2 articles, maybe they can be of use to you:
- "Ageing shocks and short-run regional labour market dynamics in a spatial panel VAR"
- "Ripple effects of house prices: considering spatial correlations in geography and demography"
Good luck!
Caro Diogo, obrigado pelo vídeo. De grande valor para todos! Estou com um problema em relação à importação da matriz do GeoDa. Tentei esse código: spwmatrix import using "$data\Weight.gal", wname(Wcontig) xport(Wcontig, txt), mas tem dado erro (aparece a mensagem file Wcontig.txt could not be opened). Você tem alguma ideia do que possa estar acontecendo?
Hi sir, i Am a first year PhD sudent. I want to test the spatial effects of entrepreneurship on economic development for the middle East and North Africa region. Is that possible? If yes, I can't extract the shapefile for the whole region! Please I need your help
It is possible. First you have to define how you're going to measure "entrepreneurship", because that's going to be your dependent variable (and you probably will want to include "entrepreneurship in neighbouring spacial units" as an independent variable too). Your spatial units could be countries or --probably better-- cities. A distance matrix will work best in this case. It all depends on how you frame your goal. Good luck!
@ thank you sir. If you Are interested we can publish this article together. Just send me your e-mail. Otherwise I am very grateful for your help. Thank you again.
Hey Diogo. Great video
I have successfully created the weight matrix for my country. Can even run the basic reg. command.
However, whenever I try to use one of the spatial models (example), the following error occurs.
Example: xsmle dun popdens month_6, wmat(W) model(sar) fe nolog
My Panel is Strongly ballanced.
*: 3200 conformability error
_xsmle_est(): - function returned error
: - function returned error
My Stata is fully up to date and I literally have tried every single link on google.
Would be thankful for any sort of advice as to what might be causing this issue.
Lukas
did you set your data as panel data?
@
Yes it is set as Panel Data. Every single step, until the actual spatial estimation ran smoothly and without any issues.
Hi: I have a problem again. I am very satisfied if you can help me. The weight matrix I use is the threshold matrix. I did not have any problem for the lower values, but when I made it for 400km it gave me an error.(Initial values not feasible R (1400))...I how to this problem..
Hi. I suggest you read what's been said in these 2 links:
1.
www.statalist.org/forums/forum/general-stata-discussion/general/1400685-problem-with-weights-matrix-in-spatial-panel-model-xsmle
2.
www.statalist.org/forums/forum/general-stata-discussion/general/1398048-xsmle-initial-values-not-feasible-numerical-overflow
Hi.. I want to guess with the gmm. Could you help in estimating the gmm for the spatial panel model in the stata program.
Hey Kübra, sorry for the late reply. The xsmle command uses the Quasi-Maximum Likelihood (QML) estimation method. Unfortunately, I'm not aware of any way to estimate spatial panel data models in Stata using Generalized Method of Moments (GMM) techniques... But using the QML method shouldn't be a problem: Soro, W. L.; Zhou, Y. and Wayoro, D. in “Crash Rates Analysis in China using a Spatial Panel Model” (2016, p.2) say that it's recognized that “the ML and QML methods are so far the most appropriate methods to fit spatial panels”.
thank you
www.statalist.org/forums/forum/general-stata-discussion/general/1349975-spregxt-new-stata-module-econometric-toolkit-to-estimate-spatial-panel-regression-models
(SPGMM) Spatial Panel Autoregressive Generalized Method of Moments Model
Hello Diogo, thank you so much for this informative tutorial! I have a question. Your data is perfectly balanced. Can your code/weights be applied to an unbalanced data set? Thank you again.
Xsmle won't run if you have unbalanced panel data. If you search the comment section you will find some people raising this issue and hopefully some helpful replies. There are work arounds. You can transform the missing data using mi. The xsmle paper I link in the description has a section showing how to do it. That said, if I were you I'd make sure that data I was using was balanced.
@ Cool! I appreciate your instant and helpful answer. Take care, sir.
so I get "initial values not feasible" in my SAR models...what does this mean and how can I fix it?
I suggest you do a google search like this: "Initial values not feasible" xsmle (with the " " so it searches exactly those terms). You can usually find the solution to any problem in forums.
Check these 2 links:
www.statalist.org/forums/forum/general-stata-discussion/general/1400685-problem-with-weights-matrix-in-spatial-panel-model-xsmle
www.statalist.org/forums/forum/general-stata-discussion/general/1398048-xsmle-initial-values-not-feasible-numerical-overflow
Good luck!
wonderful
Hi bro thanks for your useful video. But i have a problem when I ran mine. I ran step by step according to your commands with my data, my when I ran 'estimating fixed-effects model to perform Hausman test', it reminded me " _xsmle_hausman_ml(): 3200 conformability error". However, I used the same weight matrix for SDM and they were all okay. Can you please tell me what happened? Thanks
I did some google searching and your conformability error is probably because of the size of your spatial weights matrix.
Maybe you can ask Federico Belotti (he wrote the xsmle command) to help you: federico.belotti@uniroma2.it
Hi Kong Zhang and Diogo, I found the same error, 3200. Could you please share with me the solution? Thanks! sayeduitm@gmail.com
Hi... could you help me with unbalanced data please? I tried to follow the instruction using Beloti and friends paper but I stacked because when I enter this syntax, mi impute regress x1 = z, add(50) rseed(12345), it said that variable z not found. Could you please to help me with this? Thank you.
Hi. First of all I would advise you to stay away from unbalanced panel data sets.
You may be better off removing the variables which have missing values altogether.
Having said that, you are stuck because you, unlike the authors in the example which you are following, did not define the variable "z".
You don't have to define a "z" variable to proceed, though. You should use the variable that makes sense for your data.
Type in Stata: help mi impute regress
Also, check this video: ua-cam.com/video/i6SOlq0mjuc/v-deo.html
I hope it helps!
ohh ok Nuno... the problem is most of main control variables have missing values, so it I have to deal with it. Btw, thank you for the link.
Maybe there are alternative control variables which can serve you just as well :p
Notice that the multiple imputation method (mi impute regress) deals with unbalanced panels by filling in missing values with "plausible" values. That is to say, using the mi strategy you will end up with balanced data. You can only run the xsmle command with balanced data.
It may help knowing that, when you are using a .dta file in Stata and you run the command "mdesc", Stata will present a table with the number of missing values for each variable in that .dta file.
Thanks for your comment!
Thank youuuuuu so muchhh....
Hi, thanks for the video! I'm experiencing the error invalid 'GeoData' r(198); when I try to import the .gal file using this syntax.
spwmatrix import using "/Users/cameronmartin/Desktop/UK GeoData/GeoDa/Test.gal", wname (SPW) xport(SPW, txt)
Thanks
Did you find a way to make Stata find your .gal file? That's a basic syntax error. Sorry I took so long to reply friend. I didn't notice your comment til now
Hi, Diego thanks for your reply. No, I never found out how to get around this error on STATA. I ended up using a smaller dataset on R.
Is it possible for you to help me work through this on STATA?
Thanks
Try moving your Test.gal file to the desktop and run this command in Stata:
spwmatrix import using C:\Users\cameronmartin\Desktop\Test.gal, wname(SPW) xport(SPW, txt)
invert the slash
Please I Have a question ; how many observation we should have at least ?
I don't believe there is a minimum number of observations you should have.
Thank youu
Thanks you videio, I want to known where I can find the code.do file?
I have found just it blow the video. How stuipd am I ! hahaha , thanks a lot!
There's a link in the video description
How can we estimate direct, indirect and total effects in xsmle?
Hey. You just have to add the "effects" option to the command. Example:
xsmle v1 v2 v3, wmat(W) model(sdm) re vce(cluster mun) effects nolog
please I need help ! for my work I have only one regressor so when I execute the code I obtain ;"to few variables ".. should I HAve 2 variables at least ?
To run a regression you need 1 dependent variable (in the video, v1) and 1 or more independent variables / regressors (in the video, v2 and v3).
So, you only need 1 regressor, but you should have at least 2 variables (1 dependent variable and 1 regressor). Did you include the dependent variable?
The purpose of any regression is to understand how much the dependent variable can be explained by the independent variable(s). It's an attempt to predict a value of your dependent variable based on values of your independent variable(s).
Spatial regressions (xsmle) serve the same purpose, only you're also accounting for spatial dependency (based on the spatial weighs matrix).
Notice people can refer to the independent variable as the eplanatory variable / regressor / predictor variable / controlled variable / manipulated variable / input variable.
And to the dependent variable as the explained variable / response variable / predicted variable / measured variable / experimental variable / outcome variable / output variable.
yeees I have my dependant variable and my regressor and still have a problem
Can you type here the command line you're using? For example, "xsmle v1 v2, wmat(W) model(sdm) re"
To see if I can detect what's wrong...
xsmle v1 v2, wmat(W) model(sdm) re vce(cluster mun) nolog
thank you :)
I dont know why my stata say"main diagonal must contain zero elements"when i print "spmat dta W m*,normalize (row)"
The diagonal elements of the weights matrix have to be equal to 0 to exclude self neighbors! If there are "1"s in your main diagonal, there's something wrong with the matrix.
Your video is the best I've ever seen ！thank you very much indeed!!By the way ，can you send the”code.do” to 978202709@qq.com
ohhhh I see.
my geoda maybe have some problems . my weights matrix is wrong
@@jsjcptbtptp well thanks :) I sent you the email. good luck!
Hİ! I want to apply lagrange multiplier test and robust lagrange multiplier test for spatial panel data analysis. Can ı do it with stata. Can you help me find CODs.I am very satisfied if you can help me.
You can't do the LM tests using xsmle. I don't think there is a command in Stata that allows you to do those tests for spatial panel models.
You can do them in R though. If you are interested, google "splm: Spatial Panel Data Models in R"
ok! thank you very much.
Another option you have is to use the routines in Matlab to do those LM tests. Check the 2014 paper "Matlab Software for Spatial Panels"!
Thank you
helloo. i want to estimate general nesting model (elhorst). can you help me to estimate gns model?
Hi. I don't know that specific literature so I can't really help you. Having said that, at a quick glance, as far as I can tell, there isn't a single "general nesting model". Some of the "nested models" used are models that can be estimated with the xsmle command in Stata (SARAR and SDM).
Check this link: old.aecr.org/images/ImatgesArticles/2012/3/06_PETER.pdf
Thank you.
ıf my files .gwt how to read it to stata. .ı used '' spwmatrix import using d:/non.gwt,wname(non) xport(non,txt)'' but
spwmatrix_CalcSPweightM(): 3200 conformability error
The spwmatrix command does not allow you to import spatial weights from a .gtw file...
You can do it using the spmat command. Follow the instructions in this link:
www.stata.com/statalist/archive/2011-10/msg01385.html
my file name is 'non'. and it is save D disk. ı used 'spmat import ' but STATA give error 'cannot open non.GWT'. Can you help me this subject? I think I wrote wrong.Can you write this command as an example.
Notice what I did in 11:40. Try changing your working directory to Desktop before running the command. Let me know if it fixed your problem.
After you change your Working Directory to desktop, you have to let Stata know where your non.gwt file is. If your non.gwt file is in the Desktop, you run this command:
spmat import non using C:\Users\*YourUserName*\Desktop
on.gwt, geoda
thank you very much....my problem is solved.
hello,,it was a great video and project,,,please can u give the site that you found thoses data of portugal.
Thank you so much.
Hey, thanks for the nice comment! Try here: www.diva-gis.org/gdata
sorry,,,but i didn't found the shape_file with 278 municipalities!!
can you help me please.
thank you.
drive.google.com/file/d/0By5qMMVfcc4jRHJfNExjUV9pd2c/view
yes,,i have,,,when i dowloded the data from the site i get 4 shape_file,with 1,20,306,and 4260 municipalities,,,,my question was how did you get the shpa_file with 287 municipalities.
thank you.
Just watch what I do on second 2:55 of this video. What I uploaded to google drive is exactly what I used in the video... So the municipalities.shp file only has 278 municipalities
Nuno Álvares Pereira, this video is very good! I wanted to take this opportunity to ask for your help.
I have to deliver an article on the 06th, but I can not generate my weight matrix. I tried to follow your do-file, but I could not run the first command. Help me please!
The error that appears is:
spwmatrix import using C:\Users\Jordana\Desktop\Stata\BRMUE250GC_SIR.gal, wname(weights) xport(weights, txt)
J(): 3900 unable to allocate real [5300108,5300108]
spwmatrix_CalcSPweightM(): - function returned error
: - function returned error
But what is the necessary memory? My computer has 1T of storage in HD, 8G RAM, i3 processor.
TODO ANDA BIEN REVISE DE NUEVO
could you help me to find the shapefile of morocco
Try here: www.diva-gis.org/gdata
THANK YOU
when I create the file weights-file.gal , I can't open it because it's not created as a notepad file ! please what can I do
7:39 right-click on the weights_file.gal and left-click where it says "Open with..." and then choose to open it with Notepad
cool vid
LEGO JOEY thanks bro
maybe my code is wrong