Gephi Labels and Colors

Video Statistics and Information

Video
Captions Word Cloud
Reddit Comments
Captions
working with colors and labels is a little bit tricky in Guffey so this is going to be a tutorial on how to do some of the more hidden things with colors and labels what you can see here is a graph that I've loaded this graph was built from a Flickr network I searched for a particular tag and these are all of the tags that occurred on images with the search term these tags are linked if they were appearing as tags on the same image together and what you can see is that there's three clusters that appear here the term that I searched for was mouse and so these three clusters actually represent computer mice animal mice and Mickey Mouse but we can't tell which cluster is which and we only know that because it's data that I've looked at before now I've got it sized here according to betweenness centrality and that's also the color scheme so the reddish nodes have lower between this yellow have middle between this and then the one blue node has high between this what I want to look at is ways that I can adjust colors beyond what we're able to do over here with ranking and partitioning I also might want to show node labels or edge labels but there's no obvious place to do that in Guffey and the way that you find it is to click this tiny little icon down here at the bottom if you click that it brings open this tab that has all sorts of options for you I'm going to start with the global option and the thing there that you'll probably want to try is changing the background color that's going to be white by default but you can change it to any color you want on the spectrum by just clicking on the color so for example we could make a black background which some people really like I find it a little hard to look at so I'm going to change it back but you can change it to anything you want just by picking on this spectrum if you do ever end up with just a black circle it's because your slider here is all the way at the bottom and as you move it up more colors become available I'm just going to leave it at white for now there are some things that you can change about the nodes and edges to the nodes by default we'll start off as disks but you have options so you can change it to a 3d sphere some people really like this look you get these kind of metallic looking balls it can end up being a little bit cluttered with all the different shades if you're looking at a big graph but if you like the look there's the option you can also go for rectangles which is kind of an old-school network visualization but I tend to like the circles best on edges you can change a default color which I've actually already done in this graph but you can do that by clicking here and again you get that palette so you can pick edges of any color so I'll sort of pick this reddish color for now and then you see all the edge colors change we can also check this box so that the edges are the same color as the node they start at and there you can see our nodes getting a little bit darker the scale slider here allows us to make the edges bigger or thinner depending on what you're looking at and sometimes you may just want to turn the edges off you can do that by just clicking this show box here that can be useful when we're looking at networks with a lot of nodes that have some identifiable clusters and the edges just tend to add noise but we'll leave them on for now and then finally we can click the labels tab this gives us option to add labels to the nodes or the edges so let's start with the nodes we'll check this box and what you see are the nodes appear with their labels some of those we can make out world Disney and Mickey up here wildlife maybe rodent down here and not a lot down at the bottom and this is pretty common for what you're going to get by default but there's a lot of options that we have to change the way that those labels appear we can just move the slider and that'll resize them so if we make them smaller then we're able to pick out a little bit more we can see wireless desk desktop mighty so these are our computer mouse style mice down here we've got cat kitty kitten play rodent rat appearing over here the Disney terms up at the top and then some colors making up these really prominent nodes in the middle red white and black we can change that scale by dragging the slider we can also change the color either by picking off the spectrum like we did before so just do an example here it tends to be really hard to read anything other than white or black so I'm going to move that back down to black and if you want you can also change the font or the size just by clicking here and you get your standard font palette to choose from I'm going to leave that the same for now but there are a couple other things that you can do over here we've got a few menus so for the size instead of choosing fixed we can choose node size and that scales the labels to match the size of the node we can still move the slider but then everything remains proportional so we can see red white and black which represent these large nodes in the middle tend to stay the bigger labels some of these slightly more important nodes like animal and nature have slightly bigger nodes and then the small nodes have small labels we can also change the color from unique to object and that changes it from the color that we have selected to the color of the object that's describing this usually isn't a good idea because what happens is that our node label is laid over the node and so it's the same color as the note and you can't see most of the letters so that becomes a little bit of a pain but it's an option we're going to leave it at unique okay so that's that for node labels let's turn those off I'm going to set our size back to fixed and then turn on the edge labels now this data set didn't come with any edge label so I added a few in and you can see that those terms have appeared I put in a few numbers and then a few descriptions of the tags that are linked together again we have this slider that will control the size and you can change the font and the color in the same way if we change the size to be two node size it doesn't make a lot of sense for edges and in fact our terms just disappear for scaled you'll get a little bit of a difference but really fixed is probably your best option the fact is though that edge labels tend to be a mess and here we only have a few of them you can't see what edge they're on it's still confusing and in fact if we add any other information it gets even more confusing and that leads me to actually a useful piece of what's in here which is this little configure option over here this is something that is likely that you're going to need even if you won't need it for the reason that we're seeing now so if you click configure it allows you to pick what's being displayed as the text for your labels we can do that for the nodes and for the edges for the nodes right now it's sent to label and that works fine if I look in the data laboratory each node has a label and that appears but in some datasets you have you won't have a label for your notes you'll just have the ID and the ID is perfectly descriptive of the node but if you go to turn the node labels on you don't see anything if they don't have a label so if you come to this configure tab you can turn off label and turn on ID and then it'll show the ID text instead of the missing label text so if you ever go to show your node labels and nothing appears a good thing to do is come to this configure tab first for edges right now we're looking at labels and as I said I added a few of those in every edge also has an ID generally you're not going to want to show that but I'm going to show it here just to give you an idea of what a graph would look like if we turned on the labels for all the edges because every edge will have some text here if we turn that on we get this giant mess of numbers and no matter how small we make them we're not getting any really practical information here it's just generally not a good idea to have those edge labels on but if you're going to turn on anything for an edge label the weight is probably the thing you'll want to turn on under edges you can turn off ID and click weight click OK and now your weight values will show up here if you are concerned with edge weights a better way to do this is to vary the edge thickness depending on the weight and that's something that we'll talk about in a different tutorial but sometimes you really want to show the values and that's how you do it in this configure option for now I'm going to turn the edge labels off might want to turn the node labels on sighs those to the node size and this gives us a somewhat readable version of the graph labels are generally always going to make it more confusing but in a small network like this where we don't have too much overlap we can get some useful information by displaying those labels so there you go changing colors and labels in Guffey using this tiny little option that shows up at the bottom to reveal this hidden window
Info
Channel: jengolbeck
Views: 32,771
Rating: 4.8730159 out of 5
Keywords: gephi, network visualization, social network analysis
Id: DSKvHwcI75E
Channel Id: undefined
Length: 9min 22sec (562 seconds)
Published: Sat Jul 27 2013
Related Videos
Note
Please note that this website is currently a work in progress! Lots of interesting data and statistics to come.