 Perfect. Welcome, everybody. I'm super excited today. Essentially, two halves to this presentation. We're both talking about simply creating great visualizations and what you can do to make sure that your visualizations are actually saying what you hope they'll say. And we will also look at tools specifically to cover them, so in kind of approximately two halves. I am Laura Quinn. So those of you who've been around with this LSNTAP webinar series for a little while may know me. I was the founder of Idealware and the former executive director. I'm now an independent consultant. I've been out of the space for a little while, but I am back doing consulting both with nonprofits and actually doing a little bit of work with legal aid organizations. So I'm really excited to be doing this seminar. And I am a researcher by background. So data visualizations is one of the things that's very near and dear to my heart. I do a lot of work with both quantitative data. I also do a lot of work with qualitative data. So if anybody's interested in visualizations for qualitative data, that's also something that's pretty interesting to me. And I am an expert trainer for Idealware. Idealware for anybody who isn't familiar with it is a nonprofit that is dedicated to providing information about technology. So nonprofits like you guys can make smart technology decisions, tons and tons of free information, videos, reports, resources on the Idealware website. So take a look at that. All right. So what are we going to talk about today? So we're going to start by thinking about the purpose of the visualization, which is going to be a really important guidance in thinking through both what you're doing, like how you're putting it together and what tools you're using. We'll then think through eight principles of communicating through data. So beyond how you're doing it, how do you think about it? What should you keep in mind to make sure that you are, in fact, getting the type of visualization that will best tell your story, communicate, and be accurate and useful? We'll then look through actually how to choose essentially a chart type or a type of visualization. And then we'll move to tools. So thinking about Excel. Excel is actually surprisingly useful in this area. So we'll talk through what Excel can do, some of the things that have been added to Excel in the last couple of years. And then we'll move to data visualization specialists with a particular look at Tableau and Microsoft Power BI, as well as other options and other things that you can do. So that's our agenda for today. That is what I'm planning to cover, but obviously what you are hoping to hear about is of enormous interest to me. So I wanted to pause for a second here and get your thoughts quickly into the chat if you would. Just one really quick phrase about one thing you'd be really disappointed if I didn't talk about. So either something that's in that list or something that's not in that list for me just to get a sense as to where you're most interested. Go ahead and type that in. I'll pause for a second. So it looks like we've got one on where to start and another one on resources for learning more about Power BI. Fantastic. Great. Other folks? Other thoughts? All right. So I will assume that you all want to hear about exactly what I want to talk about other than that. Though obviously if you have thoughts or concerns at any point or questions at any point, definitely ask them. As Art mentions, if you have that question, almost certainly other people have that question as well. All right. So let's dive in here. So data, obviously, there's a lot of data in the world. It can be really messy. There's lots of stuff out there. And it can be hard to use, but prettiness does not necessarily make it better. Here is a, I had a really great time finding these examples for this slide. Here are two really God-awful charts. So let me just talk a little bit about why they're awful to give us a preamble into how we might be able to better visualize things. Over on the left-hand side here, we have, this is essentially a bar chart, but we have people instead of bars. And the most obvious thing that's awful in my eye here is it's not clear whether are we supposed to be thinking about the width as well as the height of the person? Is this in fact a square that we're using to represent this volume? And what are we to make of the area under the arms at that point? So is this volume, is this height, how is this in any way better than a bar? This is, to my mind, in always worse than a bar, except for possible prettiness. So prettiness, not a goal that should override usefulness. Here on the right-hand side, we've got a, this is essentially a pie chart that's been blown out to, I guess, attempt to present the pie slices better. But this is impossible to read. Like, what's the difference between the size of the pink, orange, yellow, and purple? They are, in fact, the answer is the purple is twice as big as all those other three, possible to know. So basically, these are pretty, like this one on the right, in fact, is pretty, you know, and it's pretty trippy, but that does not make it good. So the idea what we're going towards is we're going to something polished and professional that highlights what's important. So here is kind of an old one, but an always, an always classic format. This is from a, from Children's Now, which is a advocacy organization for low-income children in California, showing the relationship between race and enrollment in preschool. And you can see here, there's actually a ton of information on this chart. You can see things like, well, look, it looks pretty good. We've got more Latinos than white kids enrolled in preschool in California, at least in 2009. That seems like a great thing, but oh wait, there are so many more Latinos that there's this enormous gap. So we've got, you know, just eyeballing this, we've got like 60% white kids enrolled while we've got well less than 50, maybe 30% or so of Latino kids enrolled. So that's a, it shows us both the overall population and how many kids are enrolled. And we can certainly make this chart better. And we'll talk about, there's for instance, a lot of, there's things like the border on this is not telling us anything, the border on this key, there's a lot of white space between these, which is not necessarily adding a ton. So we'll talk about ways in which we could potentially make this easier to look at, but it's doing a really good job of visualizing things. So that's just kind of just an overview of what we're talking about when we're talking about data visualization. So let me just pause for a breath here. Any questions or thoughts or anything as you hear me cover that, that you're saying, oh, but I really want to hear you talk about this other thing. Into the chat while you're chatting, if you are. So we're just going to, just another piece of orientation is to think about what's the purpose of your visualization. Are you exploring, for instance? So have you just, are you, you have a mass of data and you're trying to figure out what trends are in the data. That is a really important purpose that you definitely will want a tool to help you with before you have polished final graphics. But it's not the only purpose you could have. So for instance, just looking at our thing, at our examples here. Over on the left-hand side you could plot a scatter plot like we have here. And you can clearly see a trend going on here. So that's exciting. That tells you something and you can then begin to delve down to that to understand what it might mean. Over on the right-hand side eyeballing that, is there a trend there? I don't know, maybe. It kind of looks like there could be a cyclical trend in there. So this is the type of thing that I, in fact, do a fair amount as a researcher to say, alright, is there, like this is clearly, there's something going on here. This, I don't know, you'd certainly have to look at this a couple of different ways in order to see what, if anything, you have here, whether this is just random noise. So that's one thing you might be doing, to explore the data. You might be formatting it for decision-making. So to basically put information in front of your organization, your staff, your board, people very close to the organization, so that they can use it to make their own decisions. And in this case, a dashboard is a really classic example here. And we're looking here at a dashboard that is, in fact, in Microsoft BI. Sorry, Power BI. So in this particular circumstance, you are potentially, you're not necessarily saying, here is what I want you to take from this. You are not necessarily spoon feeding your audience something with a story inherent in it, but instead you're saying, here is information for a bunch of stuff. So this, for instance, what we're looking at, this is actually health outcomes, but would be totally appropriate for something that you're asking staff members to look at, whatever, weekly or monthly, and they begin to get to know it, and they begin to know, you know, is 1800, 1815, is that good or bad? How should I think about that? Stuff like that. But as something that you're just putting out into the world, this is a lot of information without very much context. And so it feels a little overwhelming, just me and you looking at it. I don't know the context for this organization. So I see 2.3 is the mortality rate, and that's down 1.3. I mean, I guess I would assume that the mortality rate going down is good, but I don't actually even know that. Well, I guess it's green, so that means it's good. But is that substantial? Is that non substantial? It's really hard to know those things. The third of our three purposes for using data is to think about it to tell a story. So basically this is, so in the description of the session, it talked a lot about using it to describe what's going on at your organization and to put it in front of funders, and that would all fall squarely in the storytelling. So basically you want then to be able to clearly see a story, you have a point that you're trying to make in your visualization. So here you've got from LSE 2017 budget request, you've got pro bono cases as a percentage of total cases closed. And you have here a clear story that pro bono cases are going up as a way of closing cases. There is no real mistaking the points that is being made in this chart. And that is often what you want to do when you're making a point to an external audience, and you want to make sure that you don't try to make many points at the same time. You want to break it down. All right, so with that kind of as a lens for the whole thing, let's talk through eight different principles for communicating. So if we're trying to, so the first thing that we might be doing is exploring. And if you're exploring, honestly anything is, you can do anything you want, because you're not trying to communicate yet. So this is for the second two purposes, and especially for the third, that we're actually trying to communicate to an audience through data. So you want to define what question you're answering. So in a dashboard, you might have many questions in each little square on a dashboard. So actually just quickly to define a dashboard, because some people sometimes ask what I mean by that. So a dashboard is just a name for this kind of visualization, which includes a bunch of different things on one page. So as opposed to having just one graph, if there were three graphs on this page, this would still be a dashboard. This now has three graphs and six numbers, which in particular, like the more things you've got, the more it looks like a dashboard. So obviously at some point, you should stop putting things on your dashboard, because it's too many things. So in a dashboard, you might have multiple questions you're answering, but for any particular chart or metric or particular square in a dashboard, you should be able to say, here is the single question that I'm answering. So for instance, is the organization improving on this metric? How do they compare to last year? Are these results unusual? So it's more than probably, what's the mortality rate? Because we looked at, all right, the mortality rate is 9.7. And what does 9.7 mean? I don't know. Is that dead people? God forbid. So trying to make sure that you know what question your audience has and what it is you're answering. You then obviously need to turn to what data you have for that and think through the very best source of data you can get. So probably for most of you, we're talking about using your own data rather than pulling data, like pulling data randomly off the internet. I totally made up this graphic here, but I enjoyed it. 98.7% of all the facts on the internet are completely accurate, according to the internet. And if you do pull data, you want to include the source so other folks can look at the accuracy. If you're using your own data, you really want to think carefully through the cleanliness and the meaningfulness of your own data. And actually, I'm going to talk about that a little bit later. But that's something important not to forget, that just because you have collected your data doesn't mean that it is pristine and ready to be visualized. And you want to be super careful about combining data from different data sources into one data set. So for instance, if you are pulling information about how many people living in poverty are asking for family law services, you then don't want to pull off the internet statistics about people living in poverty and what they want in order to combine yours and them. So basically, 60% of people living in poverty are women and 70% of the people living in poverty who come to you for your services are looking for family law. You can't necessarily generalize anything between those two things because they're two different sample sets. They're two different populations of data. And that's true for the vast amount of things even if they seem far more similar. So you want to be super careful about combining data from one source and another source. All right. So you have your data. The next step here, we're basically going through a logical progression of how to start and how to think about things. So number one is you figure out what you want to answer. Number two is you figure out where the data is coming from. Number three, you experiment with ways to answer the question. So what kind of visualization might make sense? And we're going to actually dive into looking at different types of visualizations you might use in the next section. But definitely thinking about things like over what time period? Pure numbers versus percentage versus percentage change. And when you think about this stuff, you almost certainly want to be thinking about it experimentally and possibly with kind of a paper and pen in your head as opposed to relying on a tool to tell you how to visualize this data. The tools are going to be really crappy at knowing whether you need to use percentage or pure numbers or whether you need a bar chart versus a pie chart. They're just not good at that. So you really need to have a sense yourself of what might work there. And one of the easiest ways to know what might work is to kind of flip through a bunch of possibilities to kind of say, all right, well, what would this look like if I did X? What would I look like if I did Y? So let's look again. This is the same chart that we looked at before from LSC. Let's look again at this. There's choices that they've made here. This feels like a really quite straightforward graph. But in fact, there's a number of choices that were made as they decided what to do. Number one was how many years to show? And you could see here, like, for instance, I bet, I mean, it's possible that things skyrocketed and we're at 20% in 2007. That their trend doesn't continue past that. That's conceivable. But I think it's more likely that they simply decided that this is a substantial enough trend and more numbers made it just started to look like numbers as opposed to looking like more trend. Or there was a little bit of a blip that it went up for a little bit and then down again. And this was just a little cleaner to show their story. There's also... So that was a reason not to do more. You could also do less. So you could only choose to show three or four of these. There is a really nice trend. In fact, I wonder whether... People say that seven is almost an optimal number to have in short-term memory. I wonder whether that relates to the fact that there are seven bars here. It's not really necessarily a... It could be totally a coincidence. It's not necessarily a reason to use seven bars. But it's a nice... It's substantial looking without being overwhelming looking, which is what you're certainly going for. They've also chosen to graph the percentage of... So the pro bono cases as a percentage of total cases closed, which is in fact not an immediate and obvious metric to track. So for instance, you could track just pure pro bono cases. You could put this in a stacked bar, like we looked at with the Latino and the white preschool kids. So there is... If you put it that way, it might provide more data, but the story would be less pure. So for instance, if the number of total cases closed has gone up, which we'll hope that it has over time. So if that's gone up, then the number of pro bono cases will also go up. And it's kind of the trend of that, of pro bono cases, is kind of hidden in the trend of just cases going up total. So this becomes a really useful way to plot it. So all of this just representing that there are decisions to be made about your data that are important to think about and to make and not to expect that your software is going to make them for you. As you think about, okay, what should we... So you're experimenting with different formats and different ways to present data. There is actual cognitive research that will help you to know what is better or worse to do in terms of visualizations. And I feel like this is not... People tend to do things... do kinds of graphs because they like them when I feel like there is notable research that might lend you to do one thing over another. So here's some of what it is. So most people go back to, I think it was in the 70s or the 80s, Cleveland and McGill published a study on what... It is not immediately graph type by graph type, but it applies pretty directly to how... what people can most readily understand and interpret. So the first thing that they can most easily interpret is position along a scale. So they can see that this position on a scale, so this guide, this dot here, is considerably ahead of this dot here. That's something that everybody was easily able to interpret. And two is length. So next to position on a scale, the second most easily understood thing is the length. And the two of these together really would go to a bar chart. So a bar chart has both length and a position along a scale, making it particularly easy to understand. Three is the slope or direction. So this thing is going in this direction compared to this other direction. So that doesn't immediately apply to most kinds of charts. It has... You could be looking at trend lines there. I'm sorry, yes, I'll take that back. It immediately translates to a trend line or a line chart to be able to see whether things are going up or down. Four, you've got the angle. So this is pie chart type stuff. What's the difference between this angle and this angle? People are less able to understand and see that than other things. So this is a real reason why, for instance, a pie chart is less good than a bar chart. And then continuing down the scale, area. So being able to see the area of this circle compared to this circle, people have a hard time just conceptualizing the multiplicative effect here. So the fact that this circle... I didn't actually measure these. But is this circle twice as big as this one? Three times? Actually, I think it's about twice as big in diameter. So I believe this circle is about four times as big as this other one. But that could also be me just having a poor understanding of how area works, like many people. Even worse, volume. So the volume... So this is a 3D graph, God forbid, of any kind. There's virtually no reason ever to use a 3D graph. And then you've got curvature. So this is something like a donut chart, which actually takes a pie chart and with its ear ability to read the angle and makes it worse because you can't actually see the angle in it anymore. So this is important stuff. And it's important to recognize that there are real reasons for using one kind of chart compared to another kind of chart, as opposed to just kind of what seems fanciest or what seems nifty to you. All right. So we've been talking about experimenting and thinking through what's the right format. Obviously none of that means that you should take your data and push and pull it to make it say something that it doesn't actually say. You need to make sure that you represent your data appropriately. So one of the biggest ways in which people can really muck this up is with axes. And you actually see this done a fair amount with people really trying to misrepresent your data. So for instance, over here on the left-hand side, we've got the y-axis here starts at 34 and goes to 40 for no particular reason. And you can see what looks like this gigantic trend at that point when really if you start your axis at zero, you've got a kind of modest upward trend. There's no real reason to say this is a gigantic spike, assuming that at some point this was at zero, that zero is a reasonable possibility for this to be. And that's where reasonable people can start to disagree with each other as to when, like there are certainly cases in which it's warranted not to start with zero, but in most cases, unless you have a really strong reason not to, you want to start your axis with zero. Number six of eight, so we're getting to the end here. You want to make sure you tailor it to your audience. So if you are speaking to an internal audience or an audience that you know a lot about, you can really effectively communicate things with pretty potentially complicated stuff, like a dashboard or like an electoral map that has blue and red to represent Bush and Gore districts from the, this would be the 2000 and, right, so an election cycle several cycles ago. So, but if you think about it, so this dashboard is pretty overwhelming as we've talked about it for, if you're not expected to look at this fairly, this would be the type of thing you're expected to speed on and not necessarily something you're looking at once. And if you're from Australia and you're trying to make any sense of this Bush-Gore map, it doesn't really mean anything other than, well, there's more red than blue. I don't know, really, I guess it's a little coastal. It's hard to make anything out of this without context of history, without context of even what the states are and even honestly, it's hard to see the state lines in here. So it's hard to even connect this with what any states are unless you actually know that already. Number seven, you want to make it as simple as possible. So if it is possible to take off words and make it legible by itself, if it's possible to take out such horribleness as 3D and it still works, then you should. This graph at the top is the same graph, or it's the same information as the graph on the bottom. And it's just infinitely easier to read. This one at the top, you can't really make any sense of without the percentages on it. So if you took the numbers off, it wouldn't mean anything. So what's the difference between the slice of corporate income taxes in 2007 versus 2000? Very difficult to know. But if you look at it on this particular chart, then it's clear that there's a tiny incremental change. And kind of a related concept related to getting it to simplicity is to think about it. This is an information design principle that graphic designers often refer to as reducing the ink, is remove everything you can. If you can take it off, and so if it is something other than white space and in you can remove it, you should. If you can make it smaller, and it's not literally smaller, but if you can move things closer together and it reads as well or better, then you should. So things like this on the upper left-hand corner is how Excel used to... It's a really old default format for Excel spitting out graphs. It doesn't look anything nearly as terrible as this anymore, but that's what it used to look like. And so you've got all sorts of things you don't need here. You've got this gray background is adding nothing. The actual points... You could imagine circumstances in which it actually is important to pull out a particular point or two, but pulling out every point is really adding nothing here. So you've got the line around the key. You've got the line around the graph. You've got every single year. It's pretty clear what year becomes. It comes between 1994 and 1996. So simply giving some basis for those is sufficient. So you could probably get this to an even better state, an even cleaner state, but there's certainly... This is a lot cleaner than the first one. All right, so that was a look through our eight principles. I'm going to go through kind of in the thought of helping to choose an experiment with data visualization. I'm going to walk through just some chart types, some of the more simple ones and some that may be less familiar. Before I go there, I'd be interested just to stop and to hear your thoughts. If you would, if you do me the favor of just entering a few words into the box, so into the chat, of just one or two things that came as... that are interesting to you or kind of a bit of something interesting or something that you've taken from this section. Go ahead and take a minute and enter that into the chat. I begin to feel very lonely at times like this. Yeah, Meredith mentions that you can't see the chat window. Fortunately, you can enter things into it, but you can't see what other people have put in, which kind of is unfortunate. But we'll read it out so you can know. Great, so some of the things that people are saying here. Sorry, let me situate my screen. Lisa mentions using visualization to evaluate as well as to communicate. Yes, absolutely. A critical first step is thinking through yourself what the data says using, kind of thinking, using it to explore. Gail mentions that there are studies of what people perceive as the easiest to digest. Great. We've got two lauras, the first mentioned, simpler is often better. The other Laura mentions the hierarchy of charts. Yeah, so the kind of the cognitive research is really interesting. Fantastic. You guys are a quiet group. They're making me feel very lonely. All right. So choosing your visualization. So it's very worthwhile thinking through both sides of kind of polar opposites. It's very useful to think about the range of formats that you could use and to be familiar with them. It's also very useful to not try to do things that are nifty and different just because you can. There's no reason not to use a bar chart just because you're tired of bar charts. Although I feel that desire much of my career. So I use a lot of bar charts as I am working as a researcher and I would love to find an excuse to use some fancy like spider chart or something like that. Which I know you're going to talk about. But it's very rare that you really need to go beyond a lot of the basics. So basics like simple numbers. So numbers have a real place. So either illustrated numbers. So here's the Legal Aid Foundation of Los Angeles. And we looked at some dashboards that had numbers. Line charts. So line chart, if you have a trend, especially over time, then almost certainly you want a line chart. So if things are ordered and you can put the x-axis in order, then you should and you should use a line chart. If things are not ordered and you cannot do a line chart, you probably want to use a bar chart. They're usually the best way to do something that isn't a line chart. So for instance here we've got another stacked bar. This is Utah Legal Services. This is an internal report showing the timeliness of their staff in entering case reports or case outcomes reports, something like that. Names have been changed to protect the guilty. In this case you can see things like, for instance, that Randall has way more slightly overdue and really overdue and really, really overdue cases than everybody else. And it's got a line on it which in fact doesn't work. These are not ordered. So there's no inherent order between Julie, Tyler, Lea, and Randall. So this line is not actually doing anything. So it's a case in which there probably shouldn't be a line. Pie charts. There's almost nothing that a pie chart can do that a bar chart can't do better. There's, if you have a graphic in which you are, in which you could almost just use a number. So you have, for instance, just two or possibly three percentages, like this down here, 41% male and 59% female. This is a reasonable, it's basically just a graphic representation of these two numbers. So you can see, this is from actually the same visualization from UC Davis. You can see it starts to completely fall down when you, even if there were only four slices in this and all of these slices were the same, it's very difficult to visualize the difference between this 27.9 and the 36.9. Even they're pretty different. They're pretty sizeably different numbers. But it's not easy to see that the angle and the area are different sizes. And then this is just completely not useful. So if you really wanted to show this with this information, then either a bar chart or in fact a table, which I think is next would be more useful. This is nothing really, nothing but a table in a harder to read form. Speaking of, nothing wrong with a table of numbers. I think in fact they are often, they're underused. People go to something fancier when in a lot of cases, a lot of cases in which something would be simpler than this, if there were only three rows in this, for instance, that would be a good circumstance when possibly it would be a better, better for even an external report to use a table than to use a graph just because you have, you've got 12 numbers and they're big numbers and they're not necessarily easy to represent graphically, but they're pretty easy to conceptualize. This is a budget, obviously. There's a reason that budgets are shown in tables and not in graphs, which is just, it's much easier to scan and to understand the aggregate of it when there's a lot of numbers. Plots. So a scatter plot. I really like a scatter plot for showing, especially when exploring to be a scatter plot is basically, it's literally a dot in one place in the x-axis and one on the y. So here we have some measure of healthcare quality and some measure of community well-being. So a bubble adds a third measure, and in fact on this graph, it's not labeled what that measure is, but basically the bubble adds, if it's bigger, then it is showing something else. I think in this case, there's the implication that it's another measure of well-being, in some case. So you can see something like, this is a lot on one graph. I'm not sure this graph is actually really effective in telling you, I think they'd probably do better simply to take off this third thing and make a scatter plot. So you can see, for instance, that Maryland here is pretty high, or in fact, Oregon is the highest on bang for the block on all three of these measures. There's a circle on the upper right hand of this graph. Maps also can be a really powerful thing. Obviously, if you're looking at geographic areas, then maps are what you've got. So this is basically showing the change in deforestation based on the color change. So you can see places where there's been and a lot of deforestation are in red and places where there's been little are white. So it's a really powerful way to show are there relationships between places that are geographically close together. And then there's a lot of other things in the world. So there's a lot of nifty, in fact, here's what I mentioned is a spider chart. There's a lot of nifty ways to communicate things, and unless you're communicating to a very specialty niche audience, there's probably not a lot of reasons to use it, unfortunately. So I would love to be able to use nifty or things, but it's important to recognize that your goal is communicating through data. And it's unlikely that making people actually learn how to read a graph that is unfamiliar to them is going to be helpful to that. It's just a barrier between you and putting your data into the world. All right. Let's move on to tools. I'm going to press on without waiting for questions, but definitely if you have questions, enter them into the chat. I'm always interested. So let's start by thinking through Excel, because in fact, if you're wondering how to get started, what should I do, almost certainly the answer is start with Excel, mostly because we've got it already for the most part. So if you don't have it already, it's very cheap for nonprofits through TechSoup. It will do a whole bunch of things. So nearly any kind of static chart, I'll talk more about static in just a second, is possible. So it's not the easiest tool to use at all. So basically it's possible to learn how to use it to do virtually anything you might want to do. I don't believe I have ever as a researcher professionally found a graph that I couldn't create in Microsoft Excel that I really needed to create. I can certainly come up with things that I couldn't do, but not that I actually had to do. So that brings up an interesting question. What would you say is the easiest tool to use? In fact, I'm going to get to a tool to use. So there's a bunch of online kind of single-task tools that make it super easy to do like a bar chart or a line chart. One that I like that does a number of these is called Infogram. It's much better than Excel at trying to intuit what you... So basically you say, all right, here's my data and it is a bar chart. It's much better than Excel at trying to understand how that might go into a bar chart and to let you fiddle with it to actually make it into what you had envisioned in your mind. So I like Infogram for that. It's free if you publish your data publicly online, but otherwise it is $19 a month to actually get anything out of it. So to publish charts online or to get a graphic out of the system, which is a little annoying, you can make it, but you can't actually get it out. It would just be crazy if we had a screenshot tool, but just then. Yes, absolutely. No, you could certainly screenshot it. Sorry, I'm a bad person. No, no, totally. No, I totally agree. You lose some of the graphic resolution, but you could certainly screenshot it and put it up somewhere and I'm sure they're not coming after you for it. So as I'm just going back to Excel, I've got a question that I often get on this, which is, is Google Sheets comparable to Excel? And Google Sheets is actually in this particular area in the area of data visualizations extremely uncomfortable. It is way, way less powerful. And in fact, it offers very little flexibility to do much at all. So it's much more, it's a little, it's easier to do something, but then it's very difficult to do anything or impossible, to do anything other than what it wants to do. So if you want to change the colors or the space or there's only very limited options to do that. Well, the opposite is true of Excel. Basically anything in Excel is changeable. So I can change the width between these lines, the fonts, where my lines are, how this is formatted, what my colors are. Virtually everything is customizable through pretty complicated interfaces that make it difficult to know how to find them. So if you're going to be doing kind of bits and pieces of data analysis over time, it may well be worth your effort just to kind of get up to speed on how to format things in Excel. It used to be true that there were types of charts that you might really want to have that were not in Excel. Like for instance, for a while there wasn't a scatterplot, there wasn't a histogram. That is no longer the case. So like I said, there is, as far as I know, not really any graphs that I could imagine really seriously creating in a professional context that you can't create in Excel. So that of course does not mean that it is the only tool to use for everything. So when would you not use it? So in the example we just talked about, so an infogram type circumstance, you want to be able to just occasionally create really straightforward charts and not have to think a lot about them. Excel is going to make you think a lot about them in order to get them kind of reasonably polished looking. If you want to do a lot of exploration over time, so you want to be able to say, what would this look like as a bar chart? What would this look like as a scatterplot? All right, now I want to look at the percentage over this number of years. Excuse me, I'm losing my voice a little bit. Excel will allow you to do that, but it's going to be more tedious than some of the solutions that we'll look at. So that might lend itself more to something like a Tableau or a Microsoft BI that have easier to use interfaces for that type of exploration. So basically, you are going to be spending significant amounts of time in doing data analysis, and you want to be able to kind of invest in something that is going to save you time in the long run. That might imply going beyond Excel. If your data is coming from several different sources, so you have something that's coming from your your case management system, and you've got something that's coming from a spreadsheet, and you want to merge the two together and to be able to easily over time pull graphics off those. That is something that is certainly possible in Excel, but would also lend itself to, especially if you want to do that a lot or have it combine in the same way and pull the same graph every month. That is something that would lend itself to something like a Tableau or a Power BI. You want to be able to have, well, we should have those last two in the opposite order. You want to have the visualization online, and especially to be able to interact with it online. You can obviously, with Excel, you could take a screenshot of it and put it online. And you can have, there's some, with 365, there's some ways to get things online, but with nowhere near the power that either Tableau or Power BI will give you. So if your primary destination of your charts of graphs is online, then Excel is not probably going to be your best bet, and especially if you want people to be able to interact with them, so to be able to choose, to be able to slice and dice or to drill in or things like that, Excel is just going to fall back. I mean, conceivably to do some of that, but there's no real reason to try to hack it to that. So that is kind of just a quick overview of how Excel fits into this world, and the answer is that it is not a silly tool to be basing an entire data visualization scheme around. I, in fact, as a professional, do not use anything beyond Excel. I have used Tableau for specific projects, but it has not, for me, it would be a significant investment, and it hasn't been worth the investment to me to buy it. That being said, it's less of an investment to you. So let's talk about those tools. So this is Infogram, which we already talked about. Tableau. So I keep referring to it almost one breath. Tableau and Microsoft Power BI, they are two completely separate tools, but they actually are in almost exactly the same niche. So they're very similar to each other. So Tableau, in fact, the both of them, they are desktop tools that you install on your desktop, but they are particularly good at being able to publish things online, and there is an online piece of them. And they make it really easy to slice and dice and combine the data in order to create things that are pretty good-looking. So basically, what we're looking at here is the ability to see... That's not an awesome screenshot, but to be able to see pieces of data and to be able to say, all right, and this is what I want on the x-axis. Here's what I want on the y-axis. Here's what I want to be the actual values and to be able to move those around in relatively intuitive ways. For any of you who have ever been doing a lot of that in Excel, you'll know that trying to get Excel on the x-axis is bizarrely difficult. You need to have a... It's not an intuitive thing to try to get Excel to say, oh, no, show it the other way. Where it is easy to do in Tableau. So it is free for one data source. So if you are only pulling data, for instance, from... You have all your data dumped out into an Excel spreadsheet, then it is free if you are... Sorry, if you're then able to share that publicly online, which is probably not true of many of you. So if you have something that you don't care about much, making it public, then you can use it for free. But for most of you, it's going to be $58 per license for nonprofits through TechSoup. It's about $400 a license if you're not a nonprofit. So you get a notable, notable discount for going through TechSoup. So it has a lot of... So basically it allows you to slice and dice, easily explore, and then it allows you to create robust shared dashboards as well. So basically, you can then publish all of that information out online to let anybody see it without a license or be able to see it online in whatever format you like, including the ability to drill into it or to slice and dice on it in all sorts of different formats in a fairly polished way. So it's fairly easy to get stuff that is pretty polished together. It's another advantage of these tools over something like Excel, where Excel you need... If you want to get to something that looks pretty polished, you're going to need a graphic designer or graphic design sensibilities because basically anything is possible. Excel or Power BI, the stuff that it presents out of the box tends to be more polished to begin with, so you don't need to do as much finessing of it, and so you don't need as much graphic design experience to get something that looks pretty decent. This, to my eye, looks like something that came almost out of the box. Sorry, the graphic design is out of the box for Tableau. Microsoft BI is a very similar tool. It is generally considered to be not quite as easy to do the slice and dice part of the equation. So to basically be able to do your data exploration. So a little more of a learning curve there and a little more possible that you'll... These are complicated tools, and so they tend to lend themselves to circumstances where you're like, what just happened? What did I do? Undo, undo. You're more likely to get into an undo situation with Power BI. It is... It does have a tight integration into Excel as one could imagine to access. Anybody's using access. So it's a little tighter to pull data sets in, though, honestly, Tableau also has pretty tight integration for that type of stuff. It is free for the type of data visualization that you guys are likely to do. If you get very robust, bazillion user dashboards, kind of corporate type uses, you might have to pay for it, but it is free for most things that you guys might want to do with it. So in that has... So in general, what I hear is Tableau is easier to use, but Power BI is freer. So freer often in the nonprofit we're all being a big deal. It also provides robust shared dashboards. Here's the dashboard. So they're very comfortable in these functionalities. So how difficult is it to set Microsoft Power BI up to automatically pull data from, say, somebody's case management system? It's going to depend a lot on the case management system. So it's set up to be good at extracting data from things that have extraction methods. So if there is an API and you can... So for instance, if you could hook it up with one of the data mediator tools, then you could definitely, fairly straightforwardly hook it up. You're going to need data experience. So you're going to need to know what it is you're doing in order to get the right data out. So you'll need, like in the same way that if you were going to just do a complex report, you'd want to have data experience to make sure that you're actually reporting on what you think you're reporting on. So could a legal services organization pick up a programmer if their case management system has an API to get out the data, to build a dashboard for them, and what kind of a cost range would something like that look like? So my instinct is that you might even be able to do it without a... Well, so if you have an API, then you're basically what the cost you are thinking about is primarily the cost of creating the API to get the data out. And my instinct... So this is not something that I've specifically done before, but I think that you're probably talking in the, you know, the several thousand dollar range here. So two to ten thousand dollars. So the type of thing that with a number of different organizations using the same case management system, creating something like a feed to Tableau or a feed to Power BI as a shared service might be totally doable. And then once you've got... Basically once you've got the data out, the tools will ingest them without... And it'll basically be able to read, okay, here's a data set. So if there is a scheduled way to get a data set out of the case management system, so if you could say, all right, nightly, create this report in Excel and put it in this place on this file server, then you could probably, without a programmer, get either of these tools to go find that report and ingest it. So really the work is on the getting it out. Once it's out, it's fairly easy to get in. And a question here is, could you clarify what you mean by one of the data mediator tools? What does a data mediator tool... Oh, I forgot the name of it. There is... This is probably not that important to you. So there are... If you are doing... If you're doing a lot of data integration between multiple systems, there is more and more interest in being able to pull the data... To be able to extract data from one and put it into the other system. And there is a... There are two or three systems whose names I'm completely blanking on who basically just hold that data and basically allow people to put data up and take data down. It might be relevant to this, but it would be only one of the many of the multiple ways that you'd want to look into doing this. So it's not an entirely relevant piece of knowledge to that. I can certainly figure out the names of those tools and send them over. I'm sorry, that's not incredibly useful. No problem. We'll get them into the blog poster and into the comments on video later. Great. Other questions on Tableau or Power BI before I... Actually, while you're typing questions in, let me just mention that these are... I keep mentioning these two over and over again. These aren't the only tool tools like this. There's a bunch of comparable tools. So here is Plotly, which is probably the most frequently named after this, Periscope, ClickView. But they are by far the most friendly pricing-wise for nonprofits. So these are the type of tools that go for several hundred dollars per license. And so the fact that Power BI is free and Tableau is like 50 bucks-ish per license makes them much more attractive in my mind than these other tools. Let's see. We also had a question. So back on public. So Tableau is $58 per license if your data is made public. If this is... It does in fact mean it is free on the web for everyone to view. So not that I don't think it's literally easily Google-able, but that would have to be a possibility. So certainly you wouldn't want to do this with any client data. We'll talk a little bit about maps. So maps are actually fairly supportable by a couple of the tools we looked at, including InfoGram. If what you want to do is a... So this type of map here on the right is called a chloropath, sorry. This is actually both of these. They're both chloropath maps. This is not technically a heat map. A heat map is a different type of map, which has driven me insane because I date chloropath as a technical term. But anyway, so if you're looking to do this type of map, InfoGram has a reasonable amount of power here. So that's an interesting one to look into, including down to things like the county level, and I believe the zip code level. Tableau and Power BI both also have some significant mapping features. This is also an area in which people mention Google Sheets as an option, so Google Sheets, and it's a couple of different tools together, Google's API and Google Maps. To do not just... You can certainly do kind of dots on a map type thing with Google Maps, but you can do some more powerful things as well. All right, so I have just in wrapping up a couple other... Like if you want to dive down and drill deep on data visualization, some types of things you might want to think about. But I want to pause and see if you guys have questions or thoughts before I move on. Just quickly into the chat. Based on going through that, what tools are you thinking might be interesting for you? Give me a tool or two that you're thinking might be the most relevant for you, just out of curiosity. Yeah, I'm with Alison there. Info.Grammo will definitely check out. That is... Looks like a great tool. Yeah. Yeah, and we've got a couple of votes for exploring Power BI. Yes, with also the thought that it's already installed as part of Microsoft 365. That's a great point. That if you have Microsoft 365, you've already got it, so it's easy enough to explore and to see whether you like it. Also keep in mind that if you're exploring the free Tableau, it's certainly perfectly appropriate for trying to get a sense of how much you like it, is it enough easier that you want to pay for it compared to something else so that you can think of it as a free download before you buy. Alison suggests ArcGIS Online for free mapping. That's a great... So I know ArcGIS as a really quite complicated and powerful GIS, so Graphical Information System software. But I'm wondering whether ArcGIS Online is a easier to use version. Alison, I'm curious if you'll type a little bit more into that. Is that something that is... Is that a comparable to the installed version, or is that something that's way easier for someone who may not have the expertise with GIS systems? On a side note, we did a three-part series on GIS mapping about a year and a half ago with LSC that is all up on our YouTube channel. It is one of the most popular groups of videos that we've done and it breaks it down how to use tools like that and Fusion Tables very simply. So if you're interested in the mapping side of things, it's a great three-part series. I'll drop a link into the chat during a second. Great. And Alison mentions that it is not as complicated. So the online version is not as complicated as desktop, which does potentially put it into the realm of something that is worth thinking about for these more straightforward visualizations. Yes, certainly, in fact, I might have it in the coming slides, but certainly if you want to get very serious about mapping, there is definitely a limit to what you can do with these kinds of tools before you could go into actual GIS systems, so geographical information systems. This is the type of thing that you'd be able to do in these systems where you can do dramatically more powerful things in something like ArcGIS. So this is the idea of it's a data visualization as opposed to something that is a super powerful GIS representation. Alright, some other options to consider. So if you are going down a road of something that is very kind of polished and infgram, or sorry, it kind of information visualization like, then you really probably need something that will help you to do graphic design. So something like Adobe Illustrator or Photoshop. So as soon as you get to something where it really doesn't fit into typical data visualization type formats, then you're going to need something to help. And these are a powerful tool. They're not very comparatively expensive from TechSoup. They're also like $50ish, but you're going to need a graphic designer who knows how to use them because they're powerful tools and basically they'll do anything so you're going to need to know what you're doing with them. If you want to do something if you're going to do something that is if you're doing ongoing data analysis, it might be worth your while to think about statistical coding languages. So if I were doing for a living, so I do mostly qualitative research and have comparatively small data sets to work with. So which is why something like Excel works well for me. If I had consistently large data sets and were doing primarily quantitative data analysis for a living, it would probably be worth my while to learn a statistical coding language like Python or R, Stata or SPSS. So if someone who is doing this doing data analysis full-time, that's what they would use for this type of thing and they will do visualizations as well as everything else. So they allow you to explore to do statistical correlations and the graphic or the visualization as well. However, for most of us, you can see up in the upper left-hand corner, that's what has been coded here. So there's a huge learning curve here that you really need to know what it is that you're going about here. So for most of us, it's not worth the learning curve there. Sounding similar, but actually a totally different thing. If you are going to if you want to do online visualizations with a lot of power behind them, so to be able to do for instance very fancy websites that have complicated click-through infographics and stuff like that, there's a bunch of charting coding libraries that will allow you to basically say, okay, when I refresh the data this way I want the graph to appear this other way. It's going to speed up that process as opposed to having to design charts by hand in code. So there's a bunch of stuff that will help you with that. That's also very down in the weeds of what you might want to do. Perfect, and in fact there probably should have been a slide on GIS systems here, which is definitely very similarly to those if you're going down into the realm of mapping very much worth your while to look into GIS systems. All right. Great, so questions, thoughts, comments. I've got one comment from Colleen into the chat. Another tool for very basic infographics is PictoChart. It's very inexpensive for nonprofits. Yeah, there's a bunch of tools that are kind of in the realm of infogram which do various kinds of things, and I don't actually recall which kind of thing Pictogram is, but some of them will allow you to kind of take your data and put it into a simple template to make something that's kind of easily shareable and pretty. There's things that will help you. There's a lot of online stuff and in fact when we did the research for this presentation and for the article that Brian is sharing so I also did the research for that article we struggled a lot with what to do with all of these different tools and fundamentally talking to a bunch of people we decided you know like they're all like these single use things and how many of us need to do single use over and over again. So there's a lot of them that may be very useful. I'm sorry I don't mean to imply that PictoChart isn't, but trying to keep up with everything that's out there in terms of kind of single use providers did not seem useful to us. So we did not in fact try very hard. There's a lot of them and in fact there is a link in the article and I think in fact possibly in the slide over there there's a link in the article to a list of data visualization software and resources that's got like 200 odd tools. So if you want a lot there's a lot to go through. Fantastic. Brian I know you mentioned that this was a topic that you or yourself were very interested in it. Do you have questions that you're left with or kind of overall thoughts coming to the end? So the area that is really most interesting to me, you touched on a little bit but I would love to hear some more about is what is really the best way to use data visualizations when trying to convince people to make policy decisions. So for example if I'm pushing forward the need for responsive design mobile first design with a new online intake system and I want to take some dry numbers from Q that show that low income individuals have access to smartphones and that's increasing and that they don't have home computers and that type of stuff what type of visualization would you really use for that data when my goal is to convince people to make good policy decisions based on that data? Absolutely. Yes. So I would say that so a couple of questions I would ask myself is kind of knowing specifically my audience. So that sounds like that might be an internal decision maker and so knowing specifically is something that is more likely to attract their attention if it looks really good graphically or it might in fact the opposite be true. Is it more likely to attract attention if it looks very it's very just kind of Microsoft Excel it looks very just the facts. So kind of thinking through where in the spectrum you want to be of polished and really graphically appealing like if you were sharing it on Facebook then logically polished and graphically appealing is probably where you want to go. But my instinct is that for a lot of lawyers you might actually that might work against you that they might be like oh I don't care about your fancy charts the data is the data and so presenting it in a more basically you know cut and dry fashion might be better which all goes to the idea of how can you make it most useful to those that are making the decision. So I would say alright you've got a bunch of dry data from Pew what can you so what's the key piece of data and the key pieces of data that will tell the story you want to tell and kind of experiment with okay so what would this what could I do if I put this at like what does this look like as a line chart what does this look like if we have this as percentage of people in poverty versus people not living in poverty what does this look like so kind of thinking through what are the possibilities what could I do with this data and kind of thinking about them so thinking of it as a kind of creative brainstorm process and then having the winner of that creative brainstorm process be what is most what is the intersection of most useful in telling to me than telling my story so what best makes my point and then what intersects best with what they're going to want to see like if you know they're going if they're mostly going to care about whether what other people are doing then that is logically something to steer towards so basically thinking about the intersection of what they want and what you want to present in doing the persuading and then also thinking through the persuading as both a you know like if you think about like the book switch or other things like that the persuading is not only in the data itself but in the way that it's present so there's both the heart in the head is how that book describes it so it's basically there's the the data is only one piece of a multi-pronged reason why you might do something and what was the name of that book again oh switch so so just SWITCH a fantastic book by a set of brothers I think it's the Heath brothers I think it's H-E-A-T-H I might have that slightly wrong helping so it's basically talking about how to convince people of things how to convince people to switch doing one thing and do something else instead okay I dropped a link to the book in the chat excellent looks very very interesting I love the psychology of how you convince people to do things often the presentation the style the rhetoric matters more than the cold facts yep absolutely yep and it has both in fact it's a great book very readable fantastic other questions or thoughts great well thank you so much I will turn it over do take a look at the article that Brian has put in the chat that is a direct accompaniment to this session so it is a little more tool focused than what we talked about today so it has potentially a little more about tools that we did talk a lot about the tools and not as much about the actual visualization aspects and I really enjoyed presenting to you hope to see you at another session soon I will hand it back over to Brian for any closing thoughts excellent thank you so much for this presentation the number of tools that were covered and just the practical advice was great I remind people to check out our training calendar we still got about 10 trainings left this year we've got over 30 total and all of our videos will be up on the YouTube channel which is in our in the chat also I've got a direct link to the GIS tools but we've got over a hundred training videos there online and they're all available so we'll do a quick summary of that thank you once again Laura and to Idealware for putting on a great presentation today if you've got any questions please feel free to email us or connect with us online we're happy to help people learn how to use these tools put together a tutorial on one of them whatever you need we'll make that happen and make it a resource for the entire community to use fantastic thanks a lot