 ECR Wednesday webinar of 2019 hosted by Elife, the series that aims to give early career researchers a platform to discuss issues important to you and your research career. Today we will be discussing how to create effective figures. In the second half of the webinar we'll put your questions to our speaker. To ask a question you can type in the question box on the GoToWebinar functions panel or you can tweet us, we are at Elife community using the ECR Wednesday hashtag. Finally I'd just like to let you know that we have a recording webinar and we'll make it available on YouTube in the near future. Now I'll pass over to Brian to introduce today's expert speaker. Thank you Emma and hello everyone. Thank you for joining us today for our early career researcher Wednesday webinar on graphic design tips for creating effective scientific figures. My name is Brian Kent and I'm a post stock and chair of the Elife early career advisory group. I will be the moderator for today's webinar. Just a word about our host. Elife is a non-profit organization and an initiative led by scientists for scientists with the mission to improve all aspects of research communication in support of excellence in science. And they do that through open science and open technology innovation. The goal is to encourage and recognize the most responsible behaviors in science. Elife makes it a priority to support early career researchers and for the past five years has had an early career advisory group. This webinar series ECR Wednesdays is just one of the initiatives that Elife has launched to help support early career community. I would now like to welcome our speaker Dr. Anne Martin who is a professional graphic designer turned neuroscientist after working as a professional designer for six years and received her doctorate in neurobiology and anatomy from the University of Utah in 2018 and is now a postdoc at the University of Oregon. Over the next hour Dr. Martin will offer us advice on best practices for visual communication. A scientist presenting ideas and data graphically is essential for effective communication and yet we often receive little guidance or training. I am personally very grateful for Dr. Martin being here today because I find creating graphics for my manuscripts and presentations the most daunting task of communicating my ideas and results. Graphs and figures are the first thing I look to when I'm reading a paper and yet it is the part of my own work that I dread. So thank you Anne for being here. I am very much looking forward to picking up some tips. For everyone tuning in please follow us on Twitter at elifecommunity and with the hashtag ECR Wednesday. You can also submit questions using the side panel and we will do our very best to answer them during the Q&A session. I now welcome Dr. Martin to share best practices from graphic design so you can go ahead and share your screen. Okay. Good morning everyone. I'm just waiting for the link to pop up here that should allow me to share my screen. There it goes. All right. Let's see. So looks like that is working out well. Good morning everyone. I'm very happy to be here and I want to especially thank the organizers for allowing me to talk about this topic today. I also would like to thank our host Eli for allowing this webinar series to exist in the first place. So let's go ahead and get started talking about graphic design. So now in thinking about what makes an effective figure it's helpful perhaps to consider the reverse question. What makes a bad figure? And so I put out this question to Science Twitter a few months ago and asked what makes a bad figure? I received many different responses which you can see here. We have items ranging from having tons of panels of data, absent text, a figure legend that's longer than a page itself, no error bars. And so when you look at these different responses you can see that they typically fall within two different camps. We have either too much information or we have too little information. So clearly there's a balance that needs to be struck when we're designing our figures. We need to make sure that while we're not overwhelming with too much we also are providing that essential context that makes figures understandable. So today during this webinar I'm going to be talking about graphic design in three different contexts. So first I'm going to offer some insight into design thinking. Then I'm going to talk about our graphic design toolkit. So what sorts of elements of figure design can we use in order to communicate our information, our research. And then finally I'm going to offer some input on graphs and then give you some additional resources to go to. So let's get started thinking about design. Now let's think about this particular question. What is a researcher's ultimate goal in creating a figure? Now clearly we've done the science, we have all of this information, all of this data and we need to transfer it to a reader in a clear and understandable way. So perhaps the main ultimate goal is just to show the evidence for our findings in a clear manner. But I think that in searching for graphical excellence the giant of the data visualization world put it best. And he said that graphical excellence is that which gives the reader or the viewer the greatest number of ideas in the shortest time but the least ink in the smallest space. And that's a quote from Edward Tufty. And so there's a lot of information unpack here, a lot of ideas, but typically we have this thought of okay we have so much information and we need to convey it, but we know that simplicity is key. We know that the least amount of ink in space we use in order to transfer this idea that better it will be, the easier it will be to consume. So we have these two ideas of clarity and efficiency. We want things to be clear, but we need to be efficient with the way that we explain them. So how do we embrace clarity and efficiency in our figures? The first question to ask in order to do this is who is our audience? Now for instance in giving this webinar my audience is quite broad. So I have early career researchers potentially even as early as undergraduates up to perhaps even PIs. And so the information that I'm giving you today is going to be both basic and complex. So hold tight if you're feeling like information is a little bit basic, we'll dig in deep in areas. Now in building your research figure perhaps the most enlightening question you can ask is what type of journal you're going to be submitting to because you're going to frame information very differently depending upon what journal audience you're talking to. For instance if I'm submitting to a journal like hippocampus I might not need to explain where the hippocampus is within a mouse brain. That's probably something that's already known by people reading hippocampus. On the reverse if I'm submitting to science I might need to include a great deal more context of where the hippocampus is in order to clearly explain my research findings. You also need to consider individual needs and backgrounds. So for instance everyone coming to your research paper might not be looking for the same thing. You might have someone who is reading the paper trying to learn about a very particular method that's only covered in some of the figures. So perhaps they won't have followed the story from the initial figure and they won't be able to know the details that you've placed there very quickly. So in order to be able to be efficient with their time and get them to the information they need you need to consider each figure on its own as well as as the whole. Similarly we're all humans but there's a wide range in our abilities and so for instance we have young humans we have old humans sometimes that small type can be very difficult to read so make sure that your figure is accessible for all individuals. Lastly we also have different processes of review so a large proportion of the population will probably consume their media digitally but we also have individuals who will be printing out figures and perhaps even printing them out in grayscale and so being able to understand a figure even in grayscale is a critical point to address. So how do you frame your paper for your audience and even more largely how do you organize a paper in the first place. Now typically I was taught to start very broad and then get more specific over time and this is reflected in the way that our papers are typically organized. You start off with an introduction that introduces the broad topic and then you get more specific as you go into your results showing your experiments that you're using to test your hypotheses and then finally ending with your discussion which is placing those results back into a larger hole and identifying the niche that you have carved out and this needs to be reflected also in your figures. So in figure one you're basically creating that context that you're going to refer to throughout the paper. In your middle figures you're providing your specific points of evidence and then perhaps in your final figure you're showing a model or showing how your findings fit within a larger framework. In addition you also potentially have supplemental figures and these include things like controls and clarifications that you also need to fit within this figure framework. So how can we take advantage of graphic design in order to tell this story and what I'd like to really try to convince you today is that you need to build a visual system. You need to help everyone unconsciously follow along with your story by setting a series of standard conventions that you follow beginning in your first figure and then following throughout and these are things such as keeping your data ordered the same way. So for instance if I'm showing a series of graphs throughout these different papers and I'm using the same sorts of conditions throughout I'd like to keep them ordered in the same manner. So for instance if I start with a wild type then I show a head and then I show a knockout keep that same order in each of your graphs as you're going throughout the paper. Similarly you can standardize how you label data. Let's say that I'll be showing a series of immunostanes beginning in the first figure. I can set a standard by which I put the condition perhaps on the left hand side here and then use the top area to perhaps show what antibody stain or the timing that I'm using and then I can keep that standard throughout each of my figures. And then finally you'll want to use design elements consistently. So for instance here I'm showing you the design element of color and again if I was using a sample such as a wild type a head and a knockout I can use color to establish for instance that wild type is black and keep that the same throughout all of my figures. This is also helpful because it shows when you introduce something new. So if I have consistently used for instance black for wild type green for knockout if I suddenly introduce red the reader will identify that and note oh wait there's something new here I don't know what that color designates I need to make sure I understand how this is different from what I'm already familiar with and so that can be a really helpful way of not only allowing the reader to follow along quickly but to pick up small changes that might occur throughout your paper. And so I mentioned using design elements consistently but what are these design elements what graphical tools do we have in order to communicate? And so next I'm going to be going over the elements of figure design or these graphical tools. And these include items such as type selection, location, shape, size, color and line. So these are the basic tools that we have in order to communicate. Let's get started with type selection. So type typically comes in two different types ha ha or flavors we have sans serif fonts and we have serif fonts and if you're not familiar with the difference if you look at the bottom of this graphic here outlined in red is showing you what a serif is. It's basically a small ornament that's tagged on to create a different feeling within the font. And so if you look on the right here there are three different examples of these different types of fonts. You can see Times New Roman which is a serif font and Ariel and Helvetica which are common sans serif fonts. Now in science typically we use sans serif fonts because they're more legible and easier to read and so I'd like to encourage you to choose a sans serif font throughout your figures. Now I know that it's a lot of fun to pick out different typefaces, different fonts and you might have one that you particularly love like perhaps papyrus is just it just sparks such joy. But let's put that to the side for the moment and realize that first and foremost our primary goal is to communicate effectively and the best way to do that is with the most readily legible font. So let's keep it to a sans serif font for now. Also you'll notice that for several different grant applications they'll actually specify sans serif fonts like Ariel and Helvetica for this precise reason. Now there is an exception to this and this would be called Courier. This is a font that is very helpful particularly if you're using sequences. You can see here that I've lined up several different base pairs and you can see especially on this bottom line here that the letters are no longer lining up. I have several different a's and the a's are much wider than the other letters and so I'm not getting my nice it's consistently lined up base pairs. Courier however is a fixed width font and so here you can see that even though in the above font the widths are different in Courier it's kept the same and so this can be a great trick for having you're having to display a different if you're having to display sequences. Now when we start adding type to a page we need to think in terms of a hierarchy that we're creating and how we're guiding the eye along the page and that we can do that through different type sizes. So for instance you'll notice that in the header here we have the largest type size and so your eye directly goes there to find out what precisely it is that we're going to be learning about in this graph and then next you'll identify the axes at the next largest size and then finally you'll get into the actual scale and looking at the actual differences between the data points and so just by using type size we've identified in a logical flow how we want someone to consume this information and so you'll want to keep that in mind throughout your entire figure and just knowing how you're guiding the eye based off of the type size that you're employing. One thing to keep in mind particularly with graphs is that you don't want to stack type so on the right here you can see that one letter is placed each above the other such that it spells out the word however this is very confusing to read it actually takes more time to type out as well and so if you're much better off just taking the entire word and shifting it's on its axis this is much more easier to read. Next for breaks in type typically whenever we're typing out a sentence and we have one word flow over we don't think too much about it but in a figure it causes the eye to be dragged off to the side and then it has to jump back in order to catch up that next word and so it actually takes longer to consume and it also can alter the positioning of what you're working with so for instance if this was the title for a graph and we looked at the lazy fox jumped the river it might overflow the graph in such a way that causes a problem whereas if we break it instead right down the middle the lazy fox and then break jump the river it keeps the sentence nicely positioned with the graph and allows you to consume that information that much faster. I also want to caution you against light type on a dark background I think we've all attended a lecture in which someone placed a bright yellow type onto a dark blue background and after a few slides you just can't consume it anymore it just becomes too much for your eyes and so what I want to instead encourage is for you to use a dark type on a light background now there might be the occasion where for instance when you're building a model figure that you absolutely have to use light type on a dark background if this is the case I would encourage you to make it into a bolder setting so that the lines actually thicken so that it's easier to read and also you can move to using all caps instead of using both cap settings and what this will allow is for it to be more legible this is also a good trick for if you're displaying type in a very small size you can also make it bolder and make it all caps and this will help out with it being in that smaller type size I'd also like to ask you to please limit abbreviations so believe it or not these are all abbreviations that I've seen in scientific papers someone actually shortened reward to rew so think about when you're coming to this figure as Brienne mentioned it's the first thing she looks at and if you're constantly having to look up these abbreviations you're going back and forth and back and forth and it can be very frustrating trying to figure out a number of abbreviations there's a simple rule that is exists within the field that if you have to if you can google an abbreviation then it's probably safe to use but if you can't google it then try coming up with an alternate word or try finding a different solution next please don't stretch your type so we're all scientists we're all researchers and we're all presenting this data trying to show what we found find showing our findings and whenever you alter that data or even just the type surrounding it by stretching it you're causing the reader to lose faith in what you're showing and to lose trust in that you're displaying your findings accurately and so don't stretch your type because it can make someone think that you're not actually showing them the raw version of the data so just to provide an example let's see how we can improve the text on this graph I know this is a bit of an extreme case but I feel like it really makes these different points so first off your eye directly goes straight to those fluorescent colors they're very bright and eye catching next probably you look over and you see density that's stretched over here and what I typically question whenever I see this is there's a certain amount of stretching that's happening to the to this axis but yet the data doesn't quite appear to be stretched so how did they stretch one thing and not the other I'm very confused and so I'm losing the faith that this person is actually portraying their data in in a scientific manner so how can we improve this well clearly we don't want to stretch the type I've also muted the colors so that you can pay attention to what's around the graph as well so you can pay attention to that text and also we can see that the title has now been increased in size so it's the first thing that you see then we look at the axes then we look at the actual data points and we're able to consume this graph in a more logical manner so let's move on to our next graphical element and this being location whenever we read a figure typically we read it just the same way we read a book we read from left to right top bottom and so we start up here in the top left consume to the right go down consume to the right and so as we build a figure we want to employ that same method we don't want to list a above b and then jump back up to c and then d we want to keep that horizontal left to right orientation now when we're grouping data together there are many different ways that you can use positioning or location in order to group things particularly let's look at this example so we have a number of squares and circles and based on their similarities of being color and shape we're naturally dividing this top line from the second line but what happens when we put a gap between the two now we're using white space in order to identify that the items on the left are together as a group and the items on the right are together as a group and so this proximity has now taken our original groups and subdivided them then you can also employ the use of enclosures enclosures will also group your data and even though we have this gap here we're still recognizing that this line is demarcating how information is put together within a group where I see this play a distinct role for instance is in a graph like this so this is very similar to a graph that I've viewed recently in a paper and this graph has many different problems perhaps primarily being the lack of the axes the the y and x axis titles but the point of confusion especially that I ran into with a similar graph is this addition here of the plus cocaine that's in the top right I don't know what that is referring to it could refer to the color and be identifying the data that's shown in black it could be identifying a time at which this was added to the experience experiment but there's really not a framework for me to understand this but because of the proximity of that word it's causing this to be confusing I think this is meant to be time point zero and then the addition of the drug over this time but really we can't get that input from this graph and so try to think about how when you're adding this sort of information in how you can make it such that you're not introducing thoughts that you're not intending I also want to mention that items should be placed in the order that they're mentioned within the paper so for instance if you talk about item 2a in your introduction and then you talk about item 1a you'll be dinged on that because you need to talk about things in a logical fashion and you need to keep your the items in your paper ordered in the way that you mentioned them I'd also like you to pay attention to alignment and to white space so you can see the chaos that we've represented on the left here we have a bunch of squares and a bunch of circles and we can't really identify what's meant to be tied together clearly everything's just kind of scattered all over the place and there doesn't seem to be any order that's telling us what to look at first or what's significant now if I align all of these items but it's much easier to consume it's easier to see for instance the number of different circles that we have but we also don't have any particular hierarchy between these items it seems that everything's been given equal weight and the way that you can create this hierarchy is to pay attention to white space just like with type white space can allow you to guide the eye and to find particular items before you find others now for an extreme example of this this is a wild and crazy amount of information to unpack we've got type all over the place graphics all over the place and even though type might be aligned particular graphics it's just too close together and you can't distinguish it in your eye so let's calm it down and introduce some white space and now we can get a sense of where our eye is supposed to go we can tell that this comment or this caption in bold is identified with these two images at the very top we can see that there's been a title that follows above these three individual graphics that have their own descriptions and so just by changing the white space and changing the organization of how we've introduced these graphics we've made it such that we can actually consume this data in a logical fashion next let's talk about shape and size so if you're going to try to distinguish between two different ideas using two different types of shapes you have to be able to easily understand the difference between them now clearly if I were to be creating a graph and I was going to identify data points based off of shape these two would be poor choices because they're so similar especially when I reduce them in size it's going to be very difficult to tell the difference between the two if you're going to be showing data points along a graph and you want to show two different groups use shapes that are very different such as a triangle versus a circle I want to make sure to point out something about arrows so if we look at the arrow on the left we can see that there are three different points to this arrow that are all given equal weight so if I want to point to something for instance in an immunostain and I use this arrow I could be pointing to any one of three different areas and it's not clear what it would be however if I use the arrow on the right it has a very clear focal point you can see that this arrow has directionality and that if it's actually pointing at something within a figure I'll be able to tell with accuracy the one point that I'm supposed to look at so make sure that your arrows have direction you can use shape and size in order to establish focal points so let's look at this series of shapes that we have here we have one that's very large and a different shape in the center and that's the one that your eye will go to first then it will go to the circles that are around typically following their size so we consume the item in the middle then we go to the larger of the circles and then to the smaller and so employing this within your graphic can allow you to focus in on the clear point of information that you want the person to consume first and then follow by going around in sequence one trend that I've noticed lately in figures is that it's become a little bit flashy to use area to depict a difference in percentage I'd like to caution you against this so if we were to look at the very top here where we've got wild type and knockout next to these two different lines if I were to ask you what percentage is the wild type of the knockout it's fairly easy to assume that perhaps it's around 33 percent the size of the knockout now if I was to ask you that same question down here with these two different circles it's much harder in order to tell what the actual difference is between the two our eye has a much harder time distinguishing area than it does the length so for instance with these two bottom circles I don't know that I could tell you it was 30 percent or 40 percent it's very hard to get a judge on how different these two samples actually are whereas with the length you get it automatically next is perhaps everyone's favorite element to use which is color and the one that I get the most questions on and I want to emphasize to you that colors to be used as an aid and to never be relied on to communicate information and the reason why I'm saying this is illustrated on the left so we have two different lines representing two different sets of data and these lines are identified by color based on this legend here that you see on the right now if I print this in grayscale I'm probably going to lose that information and I'll lose any ability to tell what this data actually is because I can't focus on what the color designation actually is saying so what I can do instead is simply label each line with what the data actually is so now it doesn't matter the color has been taken out of the equation and while it's an aid and while when I look from graph to graph it will be helpful if I'm viewing it in a color frame I'm not reliant upon that to get the information across now what about in a context such as this so here I'm graphing just an example of the number of animals my dog found in the past year not really but it's a fun example and here I'm using a rainbow spectrum of colors is this a problem that I'm using so many colors that can be really close together and hard to distinguish for someone who is for instance colorblind well what you can see is that for each individual column the label is appearing at the bottom and so the data isn't reliant upon the color it's actually labeled in each case and so you can still understand the information even if you lose the color a place where that doesn't hold true is for a map orientation such as this now if you take a quick glance at this map of the united states what you might conclude is that wow there is something very different about the western half of the us from the eastern half of the us because the difference between this light green and this dark green is so emphasized and so it really creates this dividing line between west and east however when you actually look at the legend down here below what you'll find is that these two color points are designating values that are right next to each other so there's actually no larger difference between this light green and this dark green than there is between the green and that light blue it's just the harshness of the value change that's making you conclude that these two areas are so different when really it's a very similar result and so rainbow maps such as this can give very misleading indications of these scale differences and i'll be sharing with you in a moment a resource that particularly if you're building a map like this will allow you to build color based on value instead of based off of hue difference and what that means is value is based off of darkness of color so for instance if you have 10 percent blue 20 percent blue 30 percent blue you're increasing darkness on a scale as opposed to choosing you know a red versus a green based off of hue and so making these sorts of maps based off of value instead of actual color differences hue differences can be very helpful in making these sorts of things more easily understood you can also use commonly known color associations but you have to do it wisely for instance in this example that's a little bit of a harsh thing to understand we're typically we think of this in terms of hot is is red and cold is blue and so if we reverse that we're losing this opportunity to make a natural connection and a quick understanding of a graph or of information because we've reversed these commonly known color associations so try and have a logical reason for choosing the colors that you're using another place where we fall into this problem is with a heat map so we're very accustomed to seeing heat maps done in two diverging colors such for instance as green and red however this is very problematic for someone for instance who is colorblind because they can't distinguish those colors very easily and so heat maps such as this could be completely lost on someone who isn't able to understand color in the same isn't able to visualize color in the same manner and so this is a very cool graphic that i'm showing you from ibm in which we're showing normal color vision and then simulating the different forms of color blindness and you can see the diversity here of how you can actually read color between these different versions of color blindness and so as i mentioned i'd like to offer you a few different resources to help out with this there are many different resources for choosing color palettes this is a question that i get all the time i want to choose really good colors to show my data but where do i go in order to pick those and so a few places that you can go are ibm.com design language um c o o l o r s dot co and then also color brewer two dot org this last one in particular will show you a map and show you colors that you can choose to distinguish areas on a map very helpfully now i also mentioned that earlier you can use value in order to choose different colors and that you need to think about this logically and what i mean is for instance let's say we're trying to show differences in concentration and so we could use white to represent uh zero and then increasing with the concentration we could increase the color and value you could also use what's called a diverging scale in which you set a midpoint um and this is commonly what's done with heat maps and then in one direction you can take it in increasing intensity of one color then the other direction and increasing intensity of another color and then finally for one other color palette option there's also categorical and this is actually just labeling categories based off of different colors so for instance a rat could be represented by orange a dog represented by purple and then in order to test your figures and make sure that they are colorblind safe we have two different options here i used to always emphasize this check as a good option to go to but their server has been down quite a bit lately and so as an alternative i haven't used it personally but color oracle is another option in which you can actually upload your figure and then it will show you how that figure is perceived under different colorblind scenarios now finally here we're going to get to line which is our last graphical element to discuss and so let's look at how line affects how we can how we visualize these two different squares if we look on the left our eye naturally goes here because we have this darkened stroke around the square um it's denoting a hierarchy by getting our eye to go directly to that particular shape but you don't want to do this too much because if you have several differing line widths then you basically create chaos now your eye is going all over the place it doesn't know where to focus and you also can't identify these groups based off of line width either because the differences are so small that you can actually pick them out in a quick manner so you don't want to use these line widths excessively i would choose one or two maybe max in order to distinguish items you can also use dotted lines to show a cut out so here i have a sample uh color block on the left and i'm using a small dotted line in order to create a region of interest and then i'm increasing that on the right and i've also increased the actual dots to show that i'm zooming in in space and so you can use these sorts of changes in line width to show a difference in dimension as well i do want to caution you about using lines without fills and graphs um so let's go back to the same uh graph where i was looking at the number of animals my dog had found and you can see that it's hard to get an idea of the differences because the lines are creating an optical uh issue with being able to distinguish one column from the next even if we then employ color and try and show these differences still those lines are thin and difficult to resolve and so instead i would encourage you to use fills in your graphs so now lastly here in the last few minutes i'm just going to talk a little bit about some concerns with graphs and also provide some resources that you can go to graphs uh are a difficult element because they have to be able to stand alone i need to be able to know just from the information that's present in your graph what exactly you're comparing there are a number of critical elements that are oftentimes left out that really you need for context in order to understand a figure one of these problems is labeling consistency so for instance if you suddenly change your variables from being on the bottom for instance to suddenly being on the left it's jarring and it takes a moment in order for the reader to be able to understand what you're showing so let's keep a measurement on for instance the y-axis variables on the x-axis and keep these consistent throughout your paper again you want to label all components and don't rely on color i know this is something i've brought up before but it really is something that is a problem and has been done so many times we typically think of gfp for instance as being green and m cherry is being red or magenta but you really need to actually put those labels on there to make sure that it's understandable for the reader i'd also like you to to encourage you to keep all your scales the same this is perhaps the most difficult to achieve because i know when we're generating those plots that sometimes the scale is automatically chosen and perhaps you don't even realize that it's altered so much when you're creating your figures but it can be something that makes a figure very difficult to understand and causes the reader to have to pause and think about it more clearly or perhaps they don't even catch it so just off the bat looking at these two different graphs it seems as if the result is very similar if not the exact same but when we then look at the scale we can see that it's very different between the two which can cause you to conclude something erroneous from what you intended and finally i know this is also repetitive but also very important try to order your data in the same way throughout your paper so if you choose an order of elements that you're showing initially like gfp and cherry then venus keep those the same throughout the paper in particular this can be frustrating if you're ordering western blots and you're trying to keep the order the same with whatever western blot you're showing but try to apply this throughout your full figure and keep things ordered in the same manner and so i just like to finish up here by offering you a few further resources that you can go to for graphs there are a few different really awesome website you can visit the our graph gallery particularly if you're generating your graphs and are offers you a lot of different options and shows you how to make many different types of graphs and i know today i've been showing you basic bar graphs to generate data but that absolutely is not an endorsement of using bar graphs for everything certainly showing different data points is it's something that you should consider doing and so our graph gallery will show you how to do those different things graph pad is another great website to go to graph pad makes prism which i typically use to generate my graphs and they have a number of different tutorials to show you how to generate these also functional art and flowing data both will help you choose what types of graphs are appropriate for your data type moving on to data visualization uh edward tufty com i mentioned him before as being a real giant in the data vis world and he has many different lessons that you can go to in order to learn more about graphic design there's a new resource that recently has come out serial mentor dot com slash data vis this is by a man named claus wilkie who has put together this really wonderful resource of a lot of different data visual data visualization tools and tricks um and it's very comprehensive and gives you really great examples of successful data visualization and ones that you want to avoid uh also uh the ecr life blog at ecrlife.org there's a great blog post by dr helena jevore uh that gives you some tips and tricks for uh making better figures and uh i encourage you to go there and check it out i've also got a blog on there that talks about actually just getting started creating your figures and what are some of the initial questions that you need to answer like rgb versus cmyk what do these things actually mean so go to ecrlife.org in order to check those out for design instruction i have a blog that i've been trying to start for a little while i'm a little slow to post but hopefully i'll get better at it um and here i'm trying to show for instance how you build things in illustrator how you can use image trace as your friend um different things to that nature uh so that's at visi.com uh and then there are also three websites that offer lessons and tutorials um that are very handy there's design dot um tuts plus um that will give you a great tutorials for illustrator and photoshop and then there's also a graphic design stack exchange for questions that you run into and then finally i know many colleges and universities partner with linda.com to give you free access to lessons there so hopefully all of these resources will be helpful for you so just to finish up i'd like to say thank you so much um i've really enjoyed giving this presentation to you all and i look forward to any questions that you might have great thank you i certainly learned a lot and just a reminder for everybody um tuning in today that we will have this recorded on our youtube channel in a few weeks so if you're wanting to go back and find those resources um it will be available so if you weren't able to write them down as she was going through they are going to be available um i also have to say i'm sorry on some of your slides there were some lines that were appearing i think it was through the go-to webinar um and i was struggling whether or not to let you know because i wasn't sure there was anything you could do about it um and also i realized that this is a presentation about visual presentation and then we had these random lines so um it wasn't too bad it didn't obscure anything that we were trying to see but um we're going to try to fix that for the recording on the youtube um yeah sometimes go-to webinar i've had it where my slides shipped so sometimes it happens so i apologize for anyone who's tuning in those random lines were not supposed to be there um great so i learned a lot uh i importantly learned about courier how it's the font size that all the letters are the same size and i think for me that's going to be very helpful in the type of work that i that i do um please send in your questions on twitter using the hashtag ecr wednesday and add your questions in in the panel on the right uh and we'll have a few minutes now to go through some of them now first um as a question what's your number one tip for knowing if you've done a good job and made your figure clear enough to understand is there is there one thing that you go to initially to to decide that yeah absolutely so i think the best thing that you can do is take your figure and put it in front of someone else so uh even along the stages before you've gotten to the very end um take it and put it in front of someone who maybe they're not even in your science field but ask them what they understand from the figure so regardless of whether or not they're in your field they should be able to tell you for instance what are you even showing what are your graphs trying to depict what conclusion do they draw from the figures that you present can they actually see these differences that you're trying to point out on this immunostain um and based off of others input it'll really help inform you what you're actually communicating great um another question what are your pet peeves in figure design definitely pet peeves are whenever there's something super distracting on the page um for instance uh let's say someone wants to uh use an ornamental font like papyrus make it super huge um and uh it's distracting because it keeps me from uh quickly consuming what i'm trying to get to and also similarly if i'm not able to understand the the key point of a graph so for instance the graph that i was showing where it was missing the axes and suddenly there was this label with plus cocaine in the in the top right hand corner i had no idea what they were trying to say with that graph which is just instantly frustrating so i think it's probably very similar to to anyone reading a figure if you're not able to actually understand what the person is trying to tell you um then that's certainly frustrating yeah that is certainly frustrating um okay next question how do you get started when you're making figures what is the first thing that you do when you're about to start so when you're crafting a paper typically you're trying to organize data um from the very start and you're trying to make it logically uh have a flow um and so you're trying to group different pieces of data um how they fit together and the easiest way to do this is basically just to take out a pen and paper and sketch it out and so uh just sketch out for instance you know that you'll have an amnesia in here it'll take about that much amount of room um you know you need a certain number of graphs um sketch it on paper we'll save you a lot of time in how to create that hierarchy um before you actually even take it to a design software that's a great idea um seems kind of obvious but some of us did not think to do that it's tempting to go to that design you know software just stare at the page for a while but much easier to to sketch and be able to erase quickly yeah no that's a great idea so another question is how do you know what font size to use when you're making a figure or diagram so that when you put it into a power point presentation it's still legible right um so going between design software and power point can sometimes be very frustrating um and it can be a bit of trial and error i'd say the very best thing that you can do is uh find a way in order to hook up your computer to a projector um even better if it's the one you're actually going to be using for your presentation and see how the type is displayed and basically test out uh the size that is to your eye appearing the one that you'll want to use um and then you'll have a framework for okay i know that if i go below this it's going to be too low if i go above this it's going to overwhelm my uh titles and my other information and so being able to have that access to a projector would be number one for figuring that out um but when you're building these graphs in the first place you're creating that hierarchy of um the title versus the axes versus the actual data and so that naturally will give you um your proper proportions if you follow that logical um size variety and so too small um i'd say it's probably a judgment factor and playing with the present uh the projector will help you with that but um definitely how you see it on a page that you're viewing will give you insight into for a powerpoint great um which software do you recommend to create figures i know you had mentioned graph pad prism do you know of any free options that you recommend yeah so being trained as a graphic designer i was trained in all of the adobe programs which i know are not free um be sure to check if you're uh employed or working at a university see if they offer it for free because it is a wonderful resource but i know it is very expensive if uh you don't have access to that for free through your university so instead there is a program called inkscape which you can use which is a vector based program also just like illustrator it's very similar to illustrator um which will allow you to build figures within that sort of platform uh within a vector platform and this is mentioned in the blog uh that i wrote for ecr life but certainly when you're making your figures um try to use a vector-based program um because that will give you freedom of scalability uh because a vector-based platform isn't reliant upon pixels it's reliant upon a series of points and if you use photoshop you're going to have to have it at a very high resolution which is going to take um a great deal of memory in order to compute um the size that you'll need for printing and so illustrator or inkscape would be your friend for making a figure there's also a program called in design which is used to create uh book layouts but i would say probably stick if you're going to learn one learn illustrator because it's the most um flexible for creating a figure great now what do you think about black slide or posters with white text um the attendee says that they've been seen seeing them gain popularity and actually find them less straining to read now you had talked a little bit about the bright colors and the the light text but what about black and white sure no that's a really good question um i actually prefer for a presentation that black background um i also find it less jarring i find it more difficult for immunostains because you have a dark background on a dark background but as you mentioned that requires you to have a light colored text um and for that the difference between white and black is less jarring than for instance a fluorescent yellow and a dark blue background but you still run into that problem of being able to see that thin um type onto a darker background so you would want to use a bolder font in order to see it easier um just to be able to distinguish that uh for all individuals okay let's see oh also i want to point out um you mentioned a poster uh and for a poster printing it out on um in a dark background sometimes that can consume a great deal of ink um and if you're conscious about wanting to use less ink on your poster um then perhaps use a lighter background uh but again it is a nice contrast um to have the white on the black in that situation too um how do you feel about grid lines for bar graphs i like grid lines but only if they're very faint um if you make them too strong they'll certainly get in the way uh and it can make it difficult in order to actually understand where your data is falling on the grid if it's too strong you'll have problems distinguishing the actual data lines from the grid lines so if you make them very faint i think they can be helpful so we have time for just one last question um but if we haven't covered your questions here just remember that we're going to continue continue um the twitter chat so we'll try our best to answer anything you have posted on twitter um but a final question from one of the attendees on a multiple linear regression graph is it allowed to put regression equations on the top of each regression line so how do you feel about the equations on the on the regression lines or how would you best present it yeah so for that it really depends on the audience that you're going towards for instance um if it's an audience where you need to have that bit of information um there then find um a way in order perhaps to have it in a size that works well which you can also do is assign for instance a letter system where you can have the equations off to the side and align a letter to to draw you to what you're actually seeing um but uh it has to be a judgment as far as are you creating something that's too busy and chaotic to understand or um is it just um you know a small number that you're adding onto there um and so depending on that you might want to pull the equations over to the side that's a great tip now we before we close do you have any closing remarks no you want to leave us with um no just wanted to uh well a couple of things that you can also do is I want to encourage you whenever you're making your figures to that you really need to keep those raw data files um and keep them separate when you're making your figures so just go ahead and create a folder put all of your raw data in there um and then create a separate set that you're actually using for cropping and creating of your figures because you'll need to go back to those and in fact journals sometimes will ask you to go back to those raw images so make sure to keep them separate in a folder that's just a little tip that I know working with people and figures I can't tell you the number of times someone said oh no I cropped that and it was my only version so be sure to keep those separately um and thanks again so much for uh hosting the webinar and allowing me to share these tips with you great well thank you um and thank you to everybody who tuned in today and contributed to the discussion it was a wonderful opportunity to learn more about creating the most effective figures for research in the life sciences I know I certainly learned a lot our next ECR Wednesday webinar will be held on February 27th and the topic will be queer and LGBTQ representation in STEM so please look for that and register and that's it for today have a great day and hope to continue the the Twitter discussions thanks everybody we're done um