Speaker 1: we said we would share some best practices with the audience. If you would highlight your main recommendations, if you said these are real takeaways for anyone listening who is interested in visualizing data and better data storytelling, what would be your key recommendations and best practices?

Speaker 2: Yeah, absolutely. I'll actually start with one I've learned recently, and I did this just to an extent, but I had Alberto Cairo on my show probably two weeks ago, and he's a big name in the data visualization space. For those who have not heard of him, he wrote How Charts Lie, the Functional Art, and he's got so many books coming out, and he's written other books. But anyways, one thing he told me, and he said he took this from someone else as well, I forgot who he credited. But when you're starting out with your dashboard, one tip he had was to write the paragraph first. Write the paragraph of what you want to tell your audience. You want to tell them sales have been decreasing the past quarter, because blah, blah, blah. Write that four or five sentence paragraph first. I've heard this said in other words where they say, write your tweet first, because it's like the short few bullets that you want to share. Then use each of those sentences as a title or header for your charts. If you're thinking to show sales trends in the past few quarters, then you might want to design a line graph. Obviously, you've already analyzed the data at that point, because you had to analyze it in order to write that paragraph. But then you think about, okay, we're working with time-series data, so we probably need a line chart. We're talking about trends, so we can add a trend line, and then you design the visualization for each of those titles. What that allows you to do is when your audience receives the dashboard, they can simply read the titles of each of the charts and be able to understand what they're looking at. Then they can use the chart to actually visualize that data and see it in their mind. Other tips, and I could probably spend hours talking about formatting. I'll go through this quickly, but basically keep it as simple as possible. I think people tend to add in images and gridlines or using the standard default settings of the data visualization software that might have black gridlines or thick gridlines that can actually detract from your audience's attention, paying attention to tick marks. It's really the little things that you can tweak in your data visualization that can simplify it. Whether it's rounding your numbers up or including useful information via annotations or subtitles, it's really these tiny little changes that you can make to take your visualization from looking good which I think most of the database softwares end up looking good, but you can make it look great and actually effectively tell that story. The other large component that I focus on a lot in data visualization is the use of color. Unless you're visualizing something related to Skittles or the rainbow, we don't need to see all the colors in there. I think you can use color and it has to be very intentional. For example, the color gray, I think is probably one of the most important colors that you can use in your data visualization because you can make all of your supporting or secondary data points in the view gray, so people can actually understand what they're looking at, but then using darker or bolder colors to actually highlight and alert your audience and basically direct their attention to what you want them to look at. I can talk about color all day though, so I'll probably stop here because I don't want to take up 20 minutes of talking about color, but that's definitely something you need to keep in mind. There's color blindness that you need to take into account. There's the fact that your data visualization might be printed in black and white, because not everybody uses the color printer, so you have to keep all of these things in mind when you're visualizing data.

Speaker 1: Yeah, I agree. The analogy I use is that I think when you want to visualize data, it's useful to put yourself into a position of a journalist. For me, journalists analyze, they investigate a story, they analyze a story, and then they collect lots of data, but then they don't just dump all of the data into the newspaper. They will very carefully craft a front page of the newspaper. For me, it starts with the headline. Exactly like you said, what's the tweet? What do you actually want to communicate here? For me, this is the headline. We actually articulate what is this data telling me. Then the two other components of any useful newspaper front page is a picture, where you then visualize data. I think all the points you've made, I would completely agree with that. We need to find useful visualization of the data, make it as simple as possible. But then I also find it's really important to bring in a narrative. Just visualizing the data for me is not enough. I think when you look at a newspaper front page, you have the headline, you have a picture, but then you also have a short narrative that summarizes the key insights and puts it all into context. Do you feel that narratives have a role in dashboards?

Speaker 2: I actually suggest subtitles for adding some additional insights. For example, if you're visualizing two categories and your title is sales have gone up, and then in the subtitle you can use category 1 has increased by X percent. I add a little bit more context in the subtitles. Then additionally, if I really wanted to highlight an outlier in a scatterplot or something similar, I use annotations. I think it also depends on where your data visualization is going. Because in a lot of cases, I focused on management reporting for the most part in my prior role. In that sense, we weren't able to put too much text in there. It was just a very interactive dashboard where people would see their KPIs, profit ratios, all that stuff. In that case, I think text is helpful. But if we keep it off the screen and maybe add it when you hover over a data point, you can get more information. That's the way I would put my context and additional information there.

Speaker 1: I agree. Annotated graphs, I think, are important.