Walkthrough: Using NodeXL to Visualize Twitter Networks

[@es70]

Fantastic material, many years later. Really great delivery & explanation.

[@--StefanusLau]

I WATCH THIS ON 2021, AND IT IS STILL USEFULL. THANKS MAN

[@duke4401]

After 9 years, this video is still helpful. Thank you so much, you save my academic life!

[@dkafeltz]

Thank you so much for this video. I'm graduating in Social Comunication in Brazil and doing a work about the interactions around a famous TV Show here on Twitter. I'm using NodeXL and some of your tips really helped me, like telling in an easy way what are Clusters and how they are made by the algorithm, or why do NodeXL chooses the more recent interactions on Twitter's search. Thanks!!

[@giadacalamai2069]

Thank you very much Dr. Cook! This video was very useful! I'm currently using Node XL Pro for analysing a network for my master thesis in Communication and Media Research at the University of Fribourg (CH)! I wanted to use both Node XL Pro and Gephi to do my analysis but maybe I don't need Gephi now that I've understood Node XL Pro better, thank you!

[@seanlynch4012]

Excellent walk through, pretty much did my assignment for me :)

[@mrnNadeem]

Thanks James, you're life saver for sure. #BestTeacherIndeed

[@riseaboveomaha]

Video has been very helpful to me, as a new NodeXL user. Although I just started using NodeXL, this video has taught me a lot! Thanks for this, James! Do you have anymore walk through videos for NodeXL?

[@JamesCookUMA]

Sure do! Just browse through my video list for examples.

[@belleli7007]

Hi James! Very clear and helpful tutorial video, thank you for sharing it! You mentioned the Social Network Analysis textbook at the end of your video, do you mind if sharing the name of it? and what textbooks do you think are good for learning social network analysis? Thanks!

[@linhhien340]

Thank you so much for the video! Very detailed and clear

[@elodiesalmon9813]

Hi Dr. Cook, Thank you very much for sharing those methods ! I am an absolute beginner with NodeXL, but I have to learn using it on my own very quickly for my PhD in History. I have 2 problems in the Vertices worksheet : 1- for some vertices lines, there are no metrics at all : no degree, no clotheness centrality, no clustering coefficient... nothing ! For 592 vertices, I have all the information, but for 277 others, absolutely nothing. So it only appears 592 vertices in the "overall metrics" worksheet, instead of 869 ! I don't understand why. I found no reason : some of the "non registered" vertices are linked in edges, some of them not, they are every where in the list... I don't understand. If I copy and past all my components in a new NodeXL file, it does the same. What can I do ? 2- For filter my network, I use the dynamic filter option, which is very useful. But I don't understand the fonction of the dynamic filter column. It has been automatically fulfilled with "FAUX" or "VRAI", and I don't understand the logical here either... I can change from FAUX to VRAI and vice versa but I don't know what it means. I hope my issues are clearly explained... Thanks so much for helping, ÉS

[@richl1481]

Hello Dr. Cook, May I know whether NodeXL still supports twitter? I have the Pro version of it. And it seems like the Twitter Users Network function only imports my followers and people I follow. There is no edge between any other nodes. I clicked on "plus friends and followers". And according to the image on the right, it should be exactly what I'm looking for. May I know whether you have also encountered this? And if so, what are some good solutions? Thanks a million!

[@gilliancasten8115]

Hi there, how do you use NodeXL to simply export an account's network, and which nodes in the network are connected? (As opposed to importing data around themes in tweets). Thank you!

[@FabriceFrossard]

Thanks a lot for this tutorial, very helpful

[@aminagomri5663]

Hi James, I'm getting an error message for edges data where the Vertex 2 column is blank - your graph shows outliers so not sure why Nodexl isn't allowing me to have data to represent outliers?

[@JamesCookUMA]

Hello, Amina! I’m not quite sure exactly what’s going on with your data, but if Vertex 2 is blank, then you indeed won’t have a tie described, because you need information regarding two Vertices - Vertex 1 AND Vertex 2 - to characterize a tie. I’m not an employee of NodeXL, so I don’t know quite what might be going on with your data. However, if you would like to send me an email with some screen captures of what you’re doing or the actual NodeXL data sheet, I would be glad to take a look. My email address is [e-mail].

[@alfonsoflores9832]

Hi Dr. Cook, I was wondering if you can tell me how to filter tweets by region in Nodexl. Ex: #adidas tweets from Mexico only...? Thank you!

[@gladistigeraldia9838]

this is very helpful! thankyou

[@lenaglt]

Dr Cook, thank you for the videos. I've been trying to use NodeXl with Twitter. Downloaded the 2007 edition, also installed Excel 2007 to be compatible. Every time i try to do the Import (from either the Twitter search network or Twitter user's network) it will release the messages like " The network couldn't be otained {webException} The remote server returned an error: (410) Gone". What am i doing wrong?

[@JamesCookUMA]

Hi, Lena glt. The 2007 version of NodeXL won't work with Twitter any more because the Twitter API (the way of obtaining data from Twitter's servers) has since been updated. You need to use the newest version of NodeXL for it to work with Twitter.

[@lenaglt]

Professor Cook, thank you for the reply. I tried everything and noticed that the 2007 NodeXl version that you use here can be still used with Twitter, but only if you have Windows 7 and previous versions. It will not work with any Windows after Windows 7.

[@KubaDusza]

Thank you, that was very helpful 🙏👍

[@venkatjagadeesh8142]

thx a lot,was really informative.

[@EmmanuelNwofe]

Could the betweeness centrality node be taken as the most influential in the network?

[@giadacalamai2069]

In my opinion, one should consider the betweenness centrality but not just consider this criterion as the only one! Closeness centrality and also in/out degree values should be considered :) at least, this is what I'm doing for my master thesis.

[@dannunya959]

I like your tutorials, so I thought I would ask you to help. I am not in your class, but I am struggling with NodeXL. I have an assignment to enter THREE vertices into the program vs two. I can easily place the first two vertices in the NodeXL spreadsheet, but can't figure out how to put the third into the spreadsheet and still have the graph function show that third list. I've been struggling with it for months, and cannot find a tutorial on that, nor is the book any help. Can you please advise?

[@JamesCookUMA]

+Dan Nunya To enter three vertices into the program, simply head to the vertices tab and enter the name of each vertex in its own row in the first column. I think you may not mean three VERTICES, though, but perhaps something else. Do you mean three MODES? If you need further help, please contact me at [e-mail] and I'd be happy to discuss this in greater detail.

[@dannunya959]

sorry professor. i didnt see this post before i wrote you the email. thanks for responding so quickly. the email explains the problem more clearly.

[@ginni123456789]

so in radian 6 i get data tweet like this "i really #love #cupcake " so will #love will be in vertex1 and cupcake will be in vertex 2? what if only one hashtag is mentioned in a tweet ? or put all tweet in edge tab how i will find pair , i mean what you mean by put any value in vertex 1 and vertex 2

[@mridulgupta6188]

+Luxmi Verma I believe you are looking for topical analysis..? Is it?

[@maraharcher5731]

Hi, Dr. Cook, how do you set up nodexl to import all tweets outside the per hour limit?

[@JamesCookUMA]

Unfortunately, you don't. It's not NodeXL that sets up the limit -- it's Twitter. Twitter's API, through which NodeXL and just about every other Twitter-scraping program accesses Twitter posts, has such a limit built in to prevent resource overuse. There are ways outside of NodeXL to set up scripts that access small number of Tweets per hour over long periods of time, but that's really in the class of a Python or R solution, not a NodeXL solution.

[@maraharcher5731]

Ok, thank you so much

[@Nouridge]

I had a question. How can I extract data from Twitter if the hashtag was active more than 1 week ago. The version I have seems to be limited by only importing tweets that occurred within the last week on my specific hashtag. I would like to extract information from a meeting that happened in February. I would really appreciate your help! Thanks.

[@JamesCookUMA]

+Nour Abuhadra Unfortunately, that's due to limitations imposed by Twitter, not a limitation imposed by NodeXL. This means that there's not any easy workaround for the problem as you encounter it in NodeXL.

[@Nouridge]

+James Cook Thank you so much for your help. Do you know any softwares that can assist me in extracting information from Twitter that dates back farther? Or would this limitation occur regardless? I've seen some interesting analyses of hashtags that date back over several years (3-4 years) and I'm looking to do a similar thing. Would appreciate any advice you can give! Thanks!

[@JamesCookUMA]

Nour Abuhadra this is most easily done capturing from the Twitter API firehose in the moment (so not headed significantly backwards in time at the moment of data capture). I unfortunately don't know of any tricks to get around this Twitter API limitation looking significantly backward in time with many tweets. If you find a workaround, please let me know!

[@ginni123456789]

Hi james i want to visualise Hashtag network , again since i have data from Radian 6 how can i visualise it from tweeter user network , as theses are not giving same coloumn name which u mentioned on hashtag network visualisations - tweeter search network

[@JamesCookUMA]

Luxmi Verma Hello again, Luxmi. I'm not exactly sure what you're asking, but I *think* you're asking whether it is possible to create a network in which hashtags are nodes (and ties indicate the frequency of co-occurrence) from data in which you have a set of users' tweets. If that is the question, the answer is a definite YES! What you would do is enter those tweets into a column in the Edges tab of NodeXL (you can put any value in for Vertex 1 and Vertex 2 for these data). Then use the command NodeXL->Graph Metrics->Words and Word Pairs to create a semantic network as seen here: kzbin.info/www/bejne/opPFY5-DeLiabJY . I hope this helps!

[@meredianalam9329]

Hi James! Thanks for the inspiring information. Anyway, I am currently researching the network of environmental movement in Indonesia with #SaveBabakanSiliwangi, but only 1 vertice appeared as the result. The vertices are supposed to be a lot because #SaveBabakan Siliwangi have been twitted by around thousand users. Can you give me a hand for this issue? Thank you very much.

[@JamesCookUMA]

Hello, Meredian. What I notice, looking at Twitter, is that the hashtag you mention has been used just once since February, with a Tweet on April 1. When you search through a hashtag using NodeXL, the Twitter API will only return material going a week or two back in time, which is why the earlier February tweets don't appear in your search results. That's an unfortunate limitation of how Twitter chooses to share its data.

[@robinnilsson9487]

I accidentally x'd out of the graph. How do I make it reappear?

[@JamesCookUMA]

Try just closing down and reopening your NodeXL file.

[@robinnilsson9487]

@@JamesCookUMA it didn't work. I figured it out though! Went to the "view" ribbon and selected the "document actions" button. It doesnt sound related at all but it worked!

[@ginni123456789]

Hi James , i have data downloaded from Radian 6, how can i import that in node xl

[@JamesCookUMA]

Luxmi Verma that depends on the form of data you have. Can you post an example of a few lines of text to indicate what your lines of data look like?

[@ginni123456789]

James Cook the data format is on excel , is there any email id where can i send the data example, i get the edge list from there but to find vertices i have been told to put the data in R software to develop relations as corresponding tweets are mentioned wondering what to do?

[@JamesCookUMA]

Excel is the file format. Edge list would be the data format, and if you have an edge list you should be able to cut and paste it in right from Excel, since the NodeXL data format is also an edge list.

[@ginni123456789]

James Cook let me explain so i have author names in one column - node but in another column i have their tweets where they are mentioning / reply to other people , if i am just copying the nodes and psting its is giving me a self loop graph how do i show relation ship here? do i need to manually from the tweet take out themention/reply twitter handle and put on vertices column to build or i have to pass it through R or any other software!! before pasting it to node xl

[@JamesCookUMA]

Luxmi Verma Oh, OK, so you *don't* have an edge list yet. You need to convert your data into an edge list, though. You certainly don't have to pass anything through R, although you could. Try this: you could expand the "Tweets" column (using the Data->Text to Columns command) so that the Tweet is separated out into different columns with a space as a delimiter. Then take every result that contains a "@" and you'll have the nodes to which they are posting -- the target. You could create your edge list that way.