hello everyone so in this video I'm going to show you the bibliographic coupling using the bibliometric software in a bit more detail in some earlier videos we have shown you how to do the bibliographic coupling sucko citation analyzes all these things using the Buell imitate but very briefly and then we also have made some videos on be blue shiny please check it out if you don't like to code much that check out our videos on people your shiny but one of the issue is blue shiny is that you do not have the bibliographic coupling option in business and you have co-citation your factorial analyzes you have social networks all these things but you don't have the bibliographic coupling and in my personal experience I find that when I want to cluster the articles I want to find different themes in the in a topic that I am researching using the bibliographic tools bibliometric tools I find the bibliographic coupling till the most robust one or one of the most robust ones okay so I really like this too but the big blue shiny tool does not support it yet so I will I will show you in a bit more detail using the course how you can do it and also I will upload this file on my website and you will find a link below in the video description so of course in our if as you already know when working with are you have to load the library first if they load the package first using the library command so which I'm doing here I have already installed it if you have not installed with the matrix place there so okay and then this command is to use the big blue shiny without the cording outlet but I'm not going to run this command for now okay because I'm going to show you with the codes how to do the bibliographic coupling so here I will load my data so my data is loaded now I have to convert my data into the record format for analyzes using the big lower tricks and we have the code here so my data is from is I and my format is plain text so the conversion is done as I can see here I have in total 279 articles so this commands here are the basic ones for some bibliometric analysis with which are not really the focus here but we can steal we can just quickly run and see what we get and you will see we get all the bibliometric information here right but you already know this if you have seen our other videos some I'm not going to spend some time here I will just go directly on the big blue metric coupling part using the codes okay and that is here on line 79 okay so here I'm creating a network matrix which I'm calling we're using that using the command bibliometric okay and amazing the converted data file M and my analysis is coupling and my network is references separated by comma semicolon okay and actually this is some other comment I don't really need it here okay so just ignore this line and it's a it's a block line anyway then this is my line for bibliographic coupling so I give it comment network plot this one this is the network that I were the plot type MDS which is multi-dimensional scaling normalize algorithm is Association clustering algorithm is walk drop and then I have here N equals 250 that is pick the most 50 most bibliographic bibliographical e coupled papers and then sighs no size is for size ax is true that means no sizes are scaled based on the number of bibliographic coupling then label is true level size is 3 and then here I'm saying level X equal to 3 it could equal to true that means level size is also scaled based on the bibliographic coupling okay and here bibliographic coupling of articles that's my title so you can actually play with different numbers difference algorithm to see all that all the options you have here let's first run this and then we'll see what are the options we have so I'm going to run this one okay and then I'll click and run this command okay so now if i zoom in here I can see by bibliographic coupled figure so here it is ok so here we have one cluster here which is like a blue one here we have red one here we have a green one here it's like violet one we have four clusters here okay but here you see the labels are so large I can't really see it nicely okay so let's say first I can just turn off the label so to do that what I will do I will just say label equals to false if I do that and run it again then you will see we don't have any labels anymore we just have the basic graph okay so that's how you can actually play with it and for labels here yeah let's let's say we keep the labels we make it shoe and then I make the size 2 1 maybe now it will be smaller and we can see it better as you can see now we can see the labels better okay but now we don't have the scaling really working well right so if I make it a little bit bigger let's say if I make it 1.5 and then I run it again it should be slightly better yeah it's slightly better okay so this looks rather okay I would say if you want to export this figure you can just click go here you can save it as made you can take it as PDF you can copy it to clipboard anyway let's explore some of the other options that we have like what are the other types that we can use what are the other normalizing algorithms we can use or what are the clustering algorithms that we can use so to do that what I will do is I will this comment I will use this command and see what are the options we have with this command I'll go for help and then start my bracket and this comment and then enter so here I will see all the options that I have with this comment okay so let's say for type here we can have Auto so it then it will do an automatic layout selection we can have circle we can have a sphere we can have MDS which I'm using here we can have first German we can have come other so let's say what happens if we make it to circle and then I run it again and if I click on plot see our plot looks like a circle we still have the same four clusters okay if you look into the colors but it looks like a circle I don't really like it much we can try something else for instance we could try the let's say we try the Kumada martyr and if we try this let's see how it looks like so normally you should play it it doesn't look well actually with this as well so maybe for this case the best option is to go with the one that I was in earlier the MDS the multi-dimensional scaling approach okay so let's say I I go with MDS for normalization we have so here you can see the details about the label level size what does the mean level kicks level color you can have also label number okay you can have yeah different algorithms so for clusters here we have Le'Veon we have HPT witness we have all these algorithms here we can have no algorithm as well so let's see if we used algorithm - another one I really recommend everyone to reading this article it already discusses most of the algorithms that they have used in the in the in this package okay so now I'm going to change the water to Le'Veon and then I will run it again let's see if our plots change a little bit it looks okay actually it looks pretty good so maybe I could use this algorithm as well and here we can see again a 1 2 3 & 4 clusters so it gives the same 4 clusters although we are using a different algorithm here for normalization we can also use a slightly different algorithm here so here we have the Association we have jacquard we have inclusion we have Sultan so let's say we use this 1000 and see what happens figure structure changed a little bit but more or less the clusters are more or less the same okay so they didn't change much so we can actually use this one as well so here this is what I want to show if you if you want to do bibliographic coupling you can also use the number of articles here - something like 60 70 whatever you want and it can change it will change based on that okay I still see I have more or less the same 4 clusters but it doesn't look really nice so one of the thing is that when doing this kind of analyzes one of the issue is that it should look nice you should have meaningful clusters okay so you have to have these things in mind here yeah I again have more or less I see that the best one I get actually with the 51 okay and here I wanted to show that you can actually change the size of the node change the size of the label you can a scale okay these informations I wanted to show you you can try the different algorithms for different layout type different normalization algorithm this different clustering algorithm so I hope this out and upload this file in in the in the research up web site link so you can download the file from there and just you you mostly have to change only the data location here and then you should be able to use all the comments it eats cake so thank you for watching if you have any questions leave it in a comment and I will try to get back to you as soon as possible and thank you for watching this [Music]