 Welcome everybody to our TechSoup event today where we get to talk about data and visualizing data. So it's gonna be a lot of fun. So we're really glad you can join us, Amelia, and share your knowledge. I'm just gonna run through a few slides here about TechSoup for anybody that may be new to our group. So I'm Sandra Amar and I'm your local TechSoup Connect host for Ontario. TechSoup Connect is a program of TechSoup and we are a global network of tech for good meetups. It's a nonprofit that helps other nonprofits to get, implement, and use technology effectively. A little bit about me, I've been working in IT for over 20 years and doing all sorts of implementations of software and projects and process improvements. And I've also worked full-time and on the side I do this and I also help smaller nonprofits and smaller businesses implement efficient digital processes and technologies to help them work more efficiently. So that's a little bit about me and I do a lot of training on Google and project management as well. So welcome again to everybody. These are a few of our community value here at TechSoup Ontario, TechSoup in general. So we welcome everyone, we put our community first. So we're here to support each other. We're all here to support each other and in order to help build stronger nonprofits and our main tool is technology as well. We would love it if you participate. We all have something to learn from each other. So again, feel free to use the chat or if you have any ideas or thoughts for other presentations, webinars or events in the future please reach out to me and let me know. And obviously we want to treat each other with kindness and respect. A little bit about TechSoup. TechSoup can help connect you to donated and discounted products as a nonprofit. You, if you sign up for TechSoup you would validate that you're a nonprofit and then you would be eligible for all sorts of discounts. These are some of the examples of some of the technology solutions and the systems that you can get at discounted and in some cases free. So for example, the Google for nonprofits is a free service but you would need to register through TechSoup and validate your status in order to get that Google for nonprofits for free. So this is an example of how much you could save based on this typical product suite. Another thing to bring to your attention if you aren't already aware is if you have any questions or you need some technical help or anything technology related for your nonprofit TechSoup also does have a forum at forums.techsoup.org and it's a great community there to ask your questions and help you find solutions to some technology issues you may be having. And with that, I am going to hand this over to Amelia to walk us through data and all the fun we can have with our data and how we can make it look pretty, which is really important because otherwise people get bored with numbers, right? Absolutely. Thanks so much, Sandra. Great to be with you all today. Okay. So my plan today is to take you on this sort of transformative journey to a version of you that makes better use of your superpowers or if you prefer your secret weapons. Okay, I'm sure that you have multiple superpowers but today I'm going to focus on just two of them which we all share, which is your ability and again, you share it with others, your staff, your board members, your funders to process visual information at lightning speed. And your second superpower is the data that your organization collects every day. Data visualization puts these two superpowers into action. It will help you to understand and show what you do and to craft new strategies to improve what you do. So here's the journey which I have visualized in a sort of graph. We begin now by discussing the mostly very rational reasons that many people in nonprofit organizations don't make good use of data. Then we'll consider your two superpowers, secret weapons in a little bit more detail. Then we'll consider your challenges and how to use those secret weapons to address those challenges. And I'll draw it to a close by pointing you to some resources that will allow you to hone your superpowers beyond this workshop. And then again, we'll open it up to questions after that. So a bit about me before we start that steep curve. I love evidence but I hate spreadsheets. I have more than 25 years of experience studying, funding and evaluating human services. Primarily as a researcher at a child policy center called Chapin Hall at the University of Chicago. At Chapin Hall, I worked with a lot of organizations and their data and came to know the difficulties that many organizations face in using data well. And I felt there had to be a better way for staff members at these organizations without high level data analysis skills to extract meaning quickly and painlessly from their data and to put it to work for them. And that's when I get interested in visualizing data, also called data viz for short. I got training in data visualization at the University of Washington, started doing it at my research center but also taking on my own clients on the side. And now I work full-time helping nonprofits, consultants and foundations to visualize data through my consulting business called data viz for nonprofits. Okay, so let's start with topic number one why you don't use data or more accurately why don't you use it as effectively as you might. I might work with organizations that come across six primary reasons why organizations don't make better use of data. So let's go through each one of them. First, there's data naivete or downright aversion. Nonprofit staff of course tend to have expertise in the issues they care about, the environment, the arts, health, education but not data analysis. Some have a downright data aversion, they admit or even probably proclaim that they aren't numbers people. Number two is time crunch. This is why we don't do a lot of things that we intend to do. We're struggling to stay afloat, to submit the next proposal, to maintain our programming of course to adjust the huge and varying needs of our clientele, cultivate new donors and digging through data just inevitably goes on the back burner. Number three is another reason we don't use data well and that's fear. Some understandably worry what their data will show. They may fear that they won't be able to control the story once their data's out there or that the data will be taken out of context that they'll be compared to other organizations that aren't truly comparable. And the worst fear is that funders will withdraw support based on data. No matter how much of those funders might talk about continuous quality improvement and such. Dirty data is another problem. Some nonprofits have low level staff entering data into management information systems or CRMs or into spreadsheets. Other have multiple staff members entering data and each staff member's entering the data a little bit differently. And the result can be what we call dirty data. So that's data that's inaccurate because it hasn't been entered in a consistent way. For example, if you have a participant in your program whose name is Michael Smith and you enter that person in once and then Michael Smith drops out and comes back six months later and gets entered into the database as Michael B Smith. Now you have dirty data because as far as the data's concerned that's two different people and the data's not gonna, the data on Michael Smith and Michael B Smith aren't gonna merge and you're not gonna be able to get a clear analysis of his trajectory. And if you have a lot of those things then your data's not gonna be very reliable. Number five is wrong data. Many nonprofits have data on their participants and their financials, but often lack data to show impact. A tutoring program, for example, may not have access to students' grades and scores because it's hard to get that data out of school systems. An employment program may not have data on their former participant's wages which would get at their possible impact. That's hard data to collect. You have to track down those past participants, survey them. So feeling like they don't have the data that they want to show sometimes organizations give up. And then finally disconnected data. So rather than having a central management information system some organizations may store their data in separate Excel or Google spreadsheets. So Michael Smith's demographics might be on one sheet and this participant's attendance in various programs might be on other sheets. And this makes analysis of, say, the relationship of age to program participation impossible because those data points are in two separate disconnected data sources. So that's another reason that we don't make better use of data. Okay, so we just talked about why you don't use data. And again, one of your secret weapons is indeed data. Most organizations have lots of data. No matter how small they are, we pack away data and databases and spreadsheets. And data is the way we know what we do is worthwhile. You can think of it as a tool and we don't invest in other tools, say computers, and then just store them away in storage rooms and rarely use them. But that's often what we do with data, right? We might skim a little bit off the top when we have a funder grant report due or something. But we also, we're often collecting way more data than we're actually using. Even small organizations collect data on many things. Here's a sample of the types of data that pretty much most nonprofits do collect. Okay, so this is the point of my presentation where I give a little commercial or public service announcement for data. Now, I understand that you probably feel that data doesn't need a commercial, it already gets overhyped. We hear and talk about data all the time. We hear about evidence-based practices and data-driven solutions and big data and key performance indicators. But in all the hype around data, I think we often lose sight of what its core importance is. So that's why I'm offering up this little public service announcement. Data is important because of how perception occurs. What we perceive is based not just on what is actually happening out there in the world, what we observe, what we might call data. But our perception is also based on what we expect to observe. So this is how it works. The brain evaluates which of a variety of probable events are actually occurring. And what we perceive is based in part on what decision the brain makes. Sure, what we perceive is also based on those incoming signals from the outside world. Again, what we might call data. But there are more signals coming from within the brain that affect our perception. Okay, so that may sound kind of confusing, but I have a great example for you and you probably have seen this before, maybe on YouTube, but maybe you're like me because you saw it as part of a presentation. So there is a two minute video. If you saw it as part of a presentation that presenter probably said, okay, I'm about to show you this two minute video. Your job is to watch this video and you're gonna see a bunch of people passing a basketball amongst themselves. And your job is to watch it and kind of many times the basketball gets passed among the participants. So if you're like me, you watched it, you dutifully counted the passes. The presenter comes back on and says, okay, raise your hand if you saw a guy in a gorilla suit walk through the frame. And if you're like me, you did not see a guy in a gorilla suit at all. Of course, the presenter shows the video again and there's no missing this guy in a gorilla suit. He walks straight through the people passing the basketball and it really gets at the power of expectations and perception. As Ani-Nas Nin said, we don't see things as they are, we see them as we are. We see them as we expect the world to be. So these inner brain signals or expectations can distort our understanding of a situation. Thus, the signals from the outside world, data are quite important to either confirming or negating our expectations. But the trick is you have to pay attention to data to make that happen. Progress in organizations and indeed in all of human history often search for the concession that our expectations might be wrong and that data might have something to teach us. You learned back in the seventh grade to form a hypothesis and then to use data to see if the hypothesis holds up. And that's really the power of data. Okay, commercial over. And I always follow it with this disclaimer that although data is part of the answer, it's not the whole answer, maybe love is, I don't know. But data is really more about questions than it is about answers. It can show you what questions you should be asking and it can help you refine those questions. It can also point you in the general right direction, but you're never going to have perfect and comprehensive data. Data will never take you all the way to your answer. You really need to combine data with your own experience and wisdom in order to make good decisions. Also, we should be wary, although not fearful of data, we should always be asking, where did the data come from? Is the source likely to be biased? If it's drawn from a sample, is that sample representative of the larger group that we're trying to understand? Why is the data missing? If there is data missing, are some survey respondents, for example, hard to reach or do they lack trust in the data collection process? We really need to understand the reliability of our data when using it. Okay, so now let's consider your second superpower and that is your visual secret weapon or superpower. So take a look at this image and think about what do you understand or comprehend first? The Mona Lisa on the left or the spreadsheet on the right? Okay, do not put your answers in the chat window. I'm gonna take a wild guess that we all understand and can get the Mona Lisa much more quickly. And that's because our visual systems have evolved over millions of years to process images essentially in parallel. So what does that mean essentially in parallel? We don't read the Mona Lisa from top to bottom and from left to, we take it all in together and understand almost instantly that this is a picture of a woman in front of a landscape, sporting a dark dress and an inscrutable smile. Words and numbers only appeared within the last few thousand years, which is really just a blip and have been consumed by most humans for actually less than a thousand years. By contrast, we and our ancestors have been processing visual information for millions of years. And while all types of sensory signals barrage, the brain especially geared for visual signals and uses about half of its volume for visual analysis. And that's because for most of human history, our survival depended on detecting dangers such as predators in the tall grass using these visual systems that we have. We appeared to be just at the beginning of our journey with data as humans, as homo sapiens. Processing words and numbers is quite a recent activity for us. Perhaps as our brains evolve, we'll be able to discern a spreadsheet at a glance. But until then, we should consider visualizing our data by translating words and numbers into visual formats that use things that we can process really quickly like color, shape, size and placement. Data visualizations greatly speed up our processing of data and so in part addresses that time crunch we talked about. DataViz also is a quick way to assess how dirty disconnected or irrelevant our data is. If the picture doesn't look right or complete, then we need to do something to improve our data. Maybe that means collecting fewer data elements but doing so more accurately. Okay, here's another little test for you. How many nines do you see in this array of numbers? Really only gonna give you like two more seconds because when I add color, it's really easy to pick out the nines right. So this is just an example how quickly we can assess visual cues like color. Here's another little test for you. Quickly try to answer these three questions using the really simple spreadsheet at the top. Were foundation of corporate grants consistently higher over time? And what were the patterns among each type of grant? Both foundation grants and corporate grants. Again, three more seconds because when I show you this exact same data in a visual format, it's way easier to digest. We can see of course that corporate grants were consistently higher over time. Foundation grants remained relatively flat with one glaring exception, that sharp increase in May. And we can even see something that's really hard to see in the spreadsheet is that corporate grants had a sort of cyclical pattern of an up-up-down pattern that repeated itself on a quarterly basis, always reaching the peak in the last month of the quarter and then declining dramatically in the first month of the next quarter. So not only is data displayed visually easier to process, it's actually appears to have more impact. So a June 2018 Washington Post article described a series of really interesting experiments in which data displayed in charts significantly reduced the misperceptions of subjects, both liberal and conservative in the United States on important political issues. So this chart shows the results of just one of those experiments. In this experiment, Republican respondents were asked whether they believed that average global temperatures were increasing, decreasing, or staying the same over the past 30 years. The respondents were divided into three groups that was control group. Those are the folks represented by the orange bars on this chart. These people received no information before being posed this question about the trend in global temperatures. They were just asked to answer the question. Another group represented by the gray bars, they were given a textual summary of the scientific consensus on global temperatures. And then they were asked the question. And the third group represented by the blue bars, they got a chart showing temperature change since 1940 as measured by four climate institutions. And as you can see, the people who saw the chart, again, they're represented by the blue bars, were the least likely to draw the wrong conclusion about global temperatures. Strangely enough, that textual summary, again, the gray bar people, that actually exacerbated the perception problem for Republicans who strongly identified with the GOP. So a chart was actually even, it was better than a textual summary, but actually a textual summary was worse than no information at all. And the results were exactly the same on the liberal side. They were given, they were asked questions that they were likely to have misperceptions of and they did best after receiving a chart too and in correcting their perceptions. Okay, so you have these secret weapons. What are they good for? Let's consider what your greatest challenges are and then we'll consider how to address those challenges using data visualization. So I'm gonna focus on four just very common challenges that organizations face, showing the need for your services, showing the impact of what you do, distinguishing your services from that of the other organizations and evaluating your own progress over time and showing that progress to various in-house and external stakeholders. Okay, so this is really the heart of the presentation. How do you use data viz to conquer the challenges that you have? And what I'm gonna do is share with you a number of visualizations that I think do a good job of addressing those challenges. One way organizations use data is to make the case for need for their services. So here's an example for you from over 200 years ago in the 1800s Florence Nightingale and a statistician named William Farr analyzed army mortality rates during the Crimean War. They discovered that most of the soldiers hadn't died in combat, but rather by preventable diseases caused by poor hygiene. Nightingale decided to make her case with pictures rather than tables of numbers or pros. In her words, quote, to effect through the eyes what we failed to convey to the public through their wordproof ears. So her invention was called the polar area chart. That's what you're looking at right now is her polar area chart. And as you can see, it's a variant on the pie chart. Each slice of the pie shows deaths for one month of the war, growing larger if the deaths increased that month. And each slice is color coded to show the cause of death. So the blue portions are deaths due to preventable causes such as poor hygiene. The red portion showed death due to wounds in battle and the black portion showed deaths due to other causes. They presented this to the Queen and to Parliament who could see it at glance, the importance of better hygiene. They quickly set up a sanitary commission to improve the conditions for the soldiers and death rates definitely fell. Nightingale became one of the first people to successfully use data visualization for persuasion to influence public policy. There's lots of other ways to show need to show the prevalence of a problem. You can use an icon chart like this one. This chart obviously uses color as well as the icons and it may more make more intuitive sense than a bar chart, at least to some viewers. You won't wanna overuse this kind of chart but our brains tend to glitch on trying to digest a bunch of percentages. And this is a good way of really making it more concrete than a number of people affected. This is called a quadrants chart and it shows the relationship between two key measures and can be another great way to show exactly where the need is. I created this chart for a client of mine who was working with a number of school districts to implement what they call deeper learning strategies. The first thing these consultants wanted to do was survey all the principals and teachers in the participating and the survey asked them questions to ascertain how important each of these deeper learning strategies were to them and how often they already put these strategies into practice. The chart shows the overall results for all principals and teachers who completed the survey. So each dot represents one of the deeper learning strategies and each dot is positioned to show the average importance assigned to the strategy and the average frequency of practice of that strategy. The consultants could see at a glance that although teachers and principals already felt that the strategies were really important, you could see they're quite high on the important scale along the access at the bottom. They were putting them into practice less than 50% of the time with the exception of direct instruction which was just over 50% of the time. So they could see the need wasn't addressing the obstacles that were preventing these educators from doing what they already believed was really important. Viewers are most interested in data about themselves. So this is a cool way to what's called put the viewer in the viz. Instead of presenting them with a whole lot of statistics that get at the need for your client's needs, the needs for your services, ask them to take a guess at how big or concerning the need is. So here's just an example. It says that 8.3% of white non-Hispanic children experienced poverty in 2019. And then it asked them to guess what the rate of poverty is among black non-Hispanic children in the same year. And you can see given a little slider, they can answer the question it's an interactive viz. And then when they scroll over those words at the bottom, they're just given this simple chart which shows them what the data actually show, what others guessed and what they guessed. So they can see how far off the mark they are. They can actually also compare themselves to others to see if they were more off the mark than others were closer to. It's a great way to put the viewer's own data into the viz and make it more engaging. Okay, let's talk about showing efficacy. Nothing's more effective than the bar chart. It may seem like a tired chart to you, but the beauty of the bar chart is we all know how to read them. So the cognitive load is really low. Another way to show efficacy is to use a line chart like this. It makes it even stronger point than a bar chart because it allows for comparison of the trajectory of one group that got an intervention to another group that didn't get an intervention. And you can see how their trajectories diverged shortly after the intervention was put into place. Still talking about efficacy, this viz uses a number of techniques which can be really helpful. The kind of chart that you're looking at is called a stacked area chart. It also uses a technique called small multiples where you place a bunch of small charts next to each other to allow for easy comparison. It also uses 100 people rather than percentages. Again, I talked about how our brains tend to glitch on percentages, but we can imagine 100 people more easily. And the notations here help too. So we can quickly see that poverty and child mortality have decreased dramatically over this time period and that basic education literacy and democracy and vaccination have all dramatically increased over the same time period. Sometimes we wanna show uniqueness. I have no idea what this kind of chart is called or whether it even has a name, but I think it shows uniqueness pretty well. It's a comparison of male and female students on advanced placement tests. So I think it effectively shows where the two groups are similar and where they diverge. So if you read it from bottom to top, you can see about the same number of male and female test takers take Latin and statistics tests and US government tests. But when you start getting into tests on art topics and languages and art history, they really diverge in terms of how many male and female test takers you have. This is called a tree map. It's also a good way to show uniqueness. It compares one organization to others using shape and color to show the number of clients. And we can very quickly see that the organization called Youth Together serves more clients than all the other organizations combine. And finally, I just want to talk about an interesting technique for using data visualization for evaluation and planning. So you may be familiar with logic models. In my work as a researcher, I came across a thousand logic models. It's just a fancy word for a flow chart. It's a flow chart showing how your organization or your program then theory is designed to work. It shows inputs and then it shows activities and then it shows the expected outputs from those activities. And finally, those outputs would relate to outcomes that you're trying to affect. But in my experience with logic models is that they get a lot of attention during the planning of a program or intervention and certainly in grant proposals, but then they collect dust when you actually put the program into place because unexpected things happen and you're too busy doing the work to see if the logic rather really played out the way it was supposed to. I thought, what happens, what would happen if we plugged a logic model into real time data? And so I came up with this idea of a living logic model. It shows progress to date using color and it's interactive. So you can see the logic models at the top and it uses color to show progress on each component of the model. So darker color show where there's been more progress, lighter color show where there's been less progress. And when you click on any one of those components, you get more data in that bar chart below. You can also scroll over components in the logic model to learn more about different aspects of the program and this sort of tooltip information can include web links to even more information. I'm happy to talk about living logic models more during Q&A. Okay, so here I'd just like to share with you 10 data vis suggestions. So in my experience, if you just follow these sort of 10 suggestions, they're not commandments, they don't apply in every single case. You can turn a good data visualization into a great one. So here you go. Suggestment number one is to clean your data first. We already talked about what dirty data is. This may seem like an obvious suggestion, but take a look at your data before you ever visualize it. Look for missing data, for duplicated data formatting issues, such as with dates. Everything rides on the quality of your data. So don't skip this step. Two is to encode your data thoughtfully. So what do I mean by that? So data encodings are white visual format you're using instead of words and numbers to show your data or using color, shape, size, placement. Researchers Cleveland and McGill have studied the types of encoding some people call them channels that people are able to decode most accurately and rank them in this on this continuum. So position along a common scale is what we can do most accurately using our human visual systems. So that's why the bar chart is so effective because we're really good at assessing the length of a bar, especially when it has a common scale. The bars all start at a common axis and then go up. We're not so good at assessing things like color hue or color intensity. We can make generic comparisons when using those types of encodings. So think about it. If I showed you two shades of green, you'd be hard pressed to say, oh yeah, this green is definitely twice as dark as that green. But if I showed you two bars, you could say pretty accurately that one bar is twice as long as the other. So when choosing encodings, we should think about where our viewers need to make the most accurate comparisons and use the correct encodings for accurate comparisons. If just more general comparisons are all that are needed, we have more options. Pies, of course, are delicious but often inscrutable when it comes to data. Looking at this image, we can confidently proclaim that the E bar is the tallest in the bar chart. But again, we're hard pressed to pick out which pie slice is largest, although these two charts show the exact same data. And the reason why the pie chart is difficult is that we're not so good at angles and pie charts involve angles. So when comparing quantities of several things, bar charts are almost always better than pie charts. I came across this eyesore of a chart recently and thought it was pretty funny that the pie chart isn't an effective way to show the numbers of countries that use different terms for this type of chart. But it's fun to know that the term for a pie chart in Catalan in Spain is cheese portions graphic and in Bulgaria, it's ceremonial bread. The only exception when a pie chart, I think is the best thing to use is when you're showing the sort of part to whole comparisons. But once you get beyond two or maybe three slices, it's usually best to skip the pie chart and use a bar chart instead or some other type of chart. Okay, so just number three is to highlight what's important. So once you visualize your data, the story that you're really trying to communicate can remain hidden, particularly those unfamiliar with the data. And that's when it's time to call out key data points using color and annotations we do in this. Here is a chart using data from the Center for Disease Control in the United States on chronic illnesses. And this chart shows an example of highlighting what's important in a chart. The type of chart is called a parallel coordinate graph and it shows the level or amount of something across several dimensions. In this case, we're looking at the percent of adults with chronic illnesses across various states. So each of those lines, all the gray lines and the one orange line represent one state in the United States. And the notations and colors draw attention to how low residents of Hawaii are in relation to other states and rates of depression. So you can see along the X-axis we have various chronic illnesses. So depression is kind of almost all the way to the right. And we can see by calling out that one line using color for Hawaii, that Hawaii is the state lowest in depression and Oregon is the state that's highest in depression. Another suggestion is to show order when it exists in your data visualization. So order data should be shown in a way that our perceptual system intrinsically senses is ordered and conversely, unordered data should not be shown in a way that perceptually implies an ordering that doesn't exist. This visualization, we're looking at a set of bar charts at the top that shows the exact same data as the set of bar charts on the bottom. But the data is health data from a health center and it shows each bar shows the percentage of health visits that included depression screening across different care teams in family medicine, pediatrics and women's health. The bars on top are not ordered. The bars on bottom are in a descending order. And you can see it just creates a lower cognitive load for the viewer. When we order those bars in descending order, we can easily see which of those care teams are meeting or exceeding goals and which are falling below those goals. Okay, clarify with color. So don't use the same color hue for two different meanings or variables. Don't use the same color saturation for different magnitudes of the same variable. Don't use too many color encodings on one visualization or data dashboard. And finally, brightly, bright, fully saturated color is like yelling and you don't wanna yell all the time. People will think you're crazy and stop listening to you. So moderate color as you might moderate your voice. Only yell when you have something really important to say and provide all the other information in a softer, less saturated color to provide contrast. So here's like the worst data dashboard in the world. It breaks all the rules I just laid out for you. It uses the same color to mean different things in side by side charts. It uses way too many colors and it uses all bright, fully saturated color. The other thing to know about colors is that we don't all see colors the same way. If these two color arrays look the same to anyone, it might mean that you have color blindness, specifically red, green color blindness. Seven to 10% of men are red, green color blind. Just fun fact, it's much lower in women because this trait exists on the X chromosome and women of course have two X chromosomes. So you would, a woman would have to have this trait on both of her X chromosomes to have red, green color blindness. But some to 10% men is rather a percentage. So we should take that into account when visualizing data. We should not be putting red and green together if that many people have trouble discerning them. These are maps showing that Center for Disease Control Chronic Illness Survey data. It uses what's called a diverging color palette to emphasize contrast. Blue tones here show where there are low number on chronic illnesses or on that overall health index. So that shows where there are healthier adults. The orange tones show where there are less healthy adults and the gray tones show somewhere in the middle. So we get that contrast. It's easy to pick out, for example, corridors of poor health in the South. Six is to delete what's unnecessary in your visualizations. So if you've read anything by Edward Tufty, who's the grandfather of modern data visualization about the data to ink ratio, which he says should be as close to one as possible. And it's really a good thing to keep in mind. Today we might think more of the data to pixel ratio rather than data to ink. So that just means eliminating anything in the visualization that doesn't communicate the data to make for a more clean look and to focus on tension on key data points. So again, I'm showing you two charts that show the exact same data, but in the one in the bottom, I've removed a lot of the non-data information or the unnecessary information. So I've removed the grid lines, the data labels, and even the color legends and just replace it with labels directly on the lines. With fewer distractions, viewers can more easily comprehend the two trends and compare the two groups more easily. The next suggestion comes from Tamara Munzer at the University of British Columbia. And that's to don't use 3D in most cases. So when we're talking about 3D, it's those cool-looking charts like an Excel that look three-dimensional that actually aren't. To make something look three-dimensional, you have to use slight foreshortening in which you reduce or distort in order to convey the illusion of three dimensions on a two-dimensional space. So parts of the image they're supposed to be perceived as closer in space are made larger. So you can see the red slice of the pie has been made larger. But as you can see, that distorts the chart because the chart on the left is the actual chart. And we can see those three pieces of the pie are exactly the same size, but the red pie slice looks bigger in the 3D chart because it had to be distorted. Similarly, to create the illusion of three dimensions, sometimes objects are placed so that they are in front, on top of each other, so they look like they're in front of each other in three-dimensional space. But of course, as you can see in these two charts, the two-dimensional version, three-dimensional version, it's really hard to see some of those bars in the green or C group in the three-dimensional version. And there's really no need for it because the data visualizers already use color to discern the three groups. So unless you absolutely need to use that 3D option, it's usually best not to use it. This one also comes from Tamara Munzer and that in her words, it's eyes beat memory, which simply means it's easy to compare things that are side by side and to hold one thing in memory and compare it and then look at something else and compare that thing to the thing you're holding in memory. So this is the power of the small multiples chart that I talked about earlier. Here's one from a great data visualizer named Doug McCune. It places several bar charts side by side. These bar charts are looking at various crimes and they're laid out together to make for easy comparisons. You can quickly scan to see the differences between different crimes. So the X axis here is time of day and the Y axis is number of crimes. Daytime crimes are represented by those yellow bars on top and nighttime crimes are represented by the blue bars on bottom. We can quickly see that driving under the influence and drunkenness occur more often during the night and that suicide and trespassing occur more often during the day. All right, zoom in. So when showing a series of visualizations, it's usually best to start with the most general and then zoom into more specific or disaggregated data. Here's an image of a series of visualizations to which you've already seen using that Center for Disease Control Chronic Illness data. The maps, when presenting the data would start with those maps which give you a general sense of where there are health concerns in the country such as that corridor for health in the Southeast. Then the parallel coordinates chart helps to compare states on specific illnesses or conditions. And then there was a last interactive dashboard which I didn't show you that allowed the viewer to dig in even deeper into trends and possible predictors of good and poor health. And my final suggestion to you is that visualization is wonderful but it's not always the answer. There are at least two instances where I think a table works better than a chart, a graph, or a map. One is when you already have a very engaged but diverse audience. So these folks are highly motivated to access certain data. They won't be annoyed by having to find their data in a table but they're diverse in their interests. Some columns or rows are gonna be of interest to some people. Different columns will be of interest to others. Tables make really efficient use of paper or screen real estate. You can get a lot of rows and columns in one table like a chart, graph, or map. So in that case, a table might be a better solution. Another instance where you might wanna use the tables when you have lots of units of measure. For example, you wanna show the height, weight, location, and satisfaction level of 100 participants in the Eating Healthy program. This data involves four different units of measure, inches, pounds, latitude, longitude, and survey ratings. Such complexity is difficult to represent on a single visualization but you can do so in a single table. So that might be another instance in which you would use table. Okay, I know we're getting close to the end. I wanna have a few minutes for questions. So I'll just go through these resources really quickly and I can share the slides with you, Sandra, people. So you don't have to be madly taking screenshots but if you're interested in checking out some different data visualization software, there's this online catalog. I'm also providing you with a list of free data visualization tools that you might wanna check out. I know that TechSoup also provides discounts to programs like Tableau and I'm happy to talk about that too about the different programs that I use. I would also suggest that if you're looking for a way to decide which charts to use for which data, that there are a number of different chart choosers online that you can use. This is probably the simplest. It's a flow chart that you just start in the middle with what you wanna show. Do you wanna show a comparison, a relationship, and then you answer several questions and it will direct you to a simple type of chart that you can use for that purpose. But there's other flow charts and chart choosing helpers out there. So I'm just showing you a number of these that you can find online. There's also this data-vis catalog that you can find along. So if you come across a chart that you think is you might wanna use but you wanna know more about it, you can look it up on this data visualization catalog. You literally just Google data visualization catalog and you'll find it. And there's a lot of data-vis gurus out there. Here are just a few that I think are really great. They all have blogs and websites. You can follow them on social media and just a little bit more about the services that I offer. I hope to be a resource to you as well. Through data-vis for nonprofits, I help nonprofits clarify the questions they wanna answer. I help them build interactive data dashboards, reports, presentations, infographics. And I do workshops like these. You should feel free to sign up for a free consultation with me on our website. I'd be happy to kick around ideas about your data with you. And I also have a blog called 60 Second Data Tips. It's a painless way to up your data game. Once a week, I send out a tip, an usually really fun tip that takes no more than 60 seconds to read. So if you just read those tips every week, you can really become much more savvy about data over time. And actually, if you put your email address in the chat window, I'm happy to sign you up for you if you don't feel like going to my website. Okay, we're two discussion. We're at the top of our curve. Now, I know we only have four minutes, but I'm happy to answer any questions that you have. Let me see. Actually, I'm gonna open the chat window. Okay. I was chatting on you. That's a great presentation. Go ahead, go ahead. Yeah, go ahead. Just clicking through the chats now. Are there any questions that I'm not seeing? I don't think I have seen any. But then I don't feel so bad taking up all the time. But if you do have questions, feel free to shoot me an email or again. Oh, there is one. Okay, good. Why do you think a tree chart is better than the much pie chart? I don't think it's always better. Again, these are suggestions, but the pie chart uses a number of things that were particularly angles that were not so good at perceiving. So I just, you can go in your own experience, but when I come across a pie chart with a whole lot of slices in it, it's really confusing. And I rarely draw much conclusion from it. Whereas if I come across the exact same data in bar chart form, I could tell you a number of things right off. I can tell you what's most, what's least, what's meeting expectations, what's below expectations if it has a reference line. So that's why I feel like other charts, not just the tree map, are often much more effective than pie chart. I think we have. Oh, I see another one about my typical fees. Okay, so the way my fees work is that I'll meet with you. I think, you know, we'll have a discussion to, for me to better understand the scope of the project. And then I will give you a flat fee. Sometimes there's a little back and forth to nail down the scope. But once the scope is nailed down, I'll give you a flat fee and if it takes me, it's based on how long I think it's going to take me to do. If it takes longer, you don't pay anymore. And my typical project is an interactive dashboard. I often create them with Tableau Public, which is the free version of Tableau. So it's short-term consultation with me. And then you have this sophisticated data dashboard that either I could tweak for you again in the future or you could tweak yourself. And I'll teach you how to refresh the data if it's not collected to a live data source. And then you're off and running. You have something to use without having to learn a whole data visualization program. For projects like that, they can range anywhere from like $1,000 to $2,000 to more like $15,000 depending on the scope of the project. Another question we had here. Any tips for visualizing qualitative data? This might be difficult to answer. Qualitative data? Yeah. Again, all these things relate to tips I have on my blog. So if you want to look at any of my past tips, I have a whole list of past tips on my website too. And this is one of them about qualitative data. So another thing that I do a lot for my clients is visualize survey data, which is often in qualitative form. And we're always looking ways for extracting more meaning from survey data. Number one is just to visualize it. I know some of our organizations that put tons of time in crafting really good surveys. And of course, the hearts part of surveys is getting a good response rate. So even organizations who are able to clear both of those hurdles and get some good usable data, then their data is never applied to action because it's really hard to extract meaning from it. So what I help them do is figure out, OK, how can we visualize this data in a way that your staff and your board members, your other stakeholders, will better understand what the implications of the data are? That often comes down to making some really good comparisons. Any data point has no meaning until it's compared to something else. So that could be comparisons over time. You want to see an increase in something or a decrease in something, comparisons with goals that you've set or benchmarks that you've set, comparisons to other organizations doing the same work. So we talk about what are the sort of informative comparisons that will set up alarm bells. Oh, this is an alert. We're really falling below with this subgroup. And with interactive data visualizations, it's really good at looking at subgroups. So maybe your overall trend is upward, but you're really falling down with this subgroup. Interactive dashboards allow you to zone in on certain subgroups really easily in a way that you can't do in a spreadsheet very easily. So when you talk about what questions you're going to need to answer at every board meeting or every staff meeting, and then customize the charts, so they're directly focused on those questions and are not going to confuse you by presenting you with a lot of information that perhaps you don't need. It looks like there's other questions coming in, but I'm going to keep it up. We have a couple more. OK. Tips on creating interactive diagrams. I'm sorry. You were breaking up. Sorry. Any tips on creating interactive diagrams? I don't know if other people can understand your voices. I'll scramble. Oh, sorry. OK. About interactive dashboards? Diagram. Again, there are a bunch of different programs out there to create interactive dashboards. The ones that the leading ones are, in my opinion, are Tableau and Power BI. Power BI is a Microsoft product. Tableau is now by Salesforce. So a lot of you might have Salesforce databases. So Tableau plays really well with Salesforce, but it also plays with Excel and tons of other databases. And it's what I use. So it's what I can talk about best. But using both those tools, you can create interactive dashboards. Data visualization tools, in my mind, exist on a continuum. At one end of the spectrum are programs like D3, which is the highest end type of data visualization program. It's what the New York Times uses. It's the most sophisticated thing out there. You have to know a lot of programming language to use it. At the way other end of the spectrum are the programs that are servled into databases that you already have. They're kind of like black boxes. The data goes in, and then it spits out a chart. And that's great, because you don't have to do anything. But the downside is that it's not very customizable. So that exact chart is not what you're looking for. You're up the creek. Tableau, and from what I understand Power BI, kind of exist in the middle of that continuum. They're drag and drop tools, so you don't have to know a ton of programming language. But they're also really sophisticated. And you can build things that are really helpful and customized to your exact purposes using programs like that. Tableau markets itself is out of the box. Start visualizing your data tomorrow. And I disagree with that characterization. It's a sophisticated tool, and there is a learning curve. So organizations have to decide, do I want to travel that learning curve, or do I want to hire somebody outside to jump start me with a data dashboard that I could then tweak into the future? I'll answer your question. Yeah, no, I think we'll wrap it up there with a couple of questions. Sandra, I can't understand you at all. Yeah, I'm not sure what's going on. I'm sorry, you're out. Sound good. Is it your Bluetooth again? Is that better? Yeah, way better. OK, I don't know what's going on. Yeah, thank you. We'll probably wrap it up here, because we are over time. There were a couple of questions in the chat, so I tried to answer. They were a little bit more generic. I tried to answer what I could. But I've also put Amelia's contact info in the chat. If you need to connect with her, let me know. And I'll share the slides, which I assume will have your contact info as well. So if anyone wants to reach out for more information, they'll know where to find you. Thank you. That was really insightful, really interesting. And I learned a few things too. I learned I'm using pie charts the wrong way. So that was great. Sandra, if you could send me whatever email I just people put in the chat, I'll sign them up for my data. Absolutely, I will. And I'll send you the slides. Excellent. Thanks for coming, everyone, and we'll see you at our next one. OK, thanks, everyone, for being with you. Bye. Bye.