Now one of the key outputs and deliverables that data analysts create are those insightful reports you present to your audience. In this data visualization module, we'll cover a little visualization theory and best practices and then we'll introduce Google Data Studio as one of the visualization tools in your toolkit for creating those actionable reports. Now, visualizing data is another key skill that's part of the data analysts skill set. The two major reasons why visualizing data is so important, one is probably pretty obvious, building those interactive deliverables and charts and graphs for your audiences is just a great tool to be able to display those insights. But often the first one that comes to mind for me as a data analyst is, if I don't know those insights off the bat and I want to explore my datasets, a lot of times a visualization tool, if you're more a visual thinker or a learner, can help uncover those insights a little bit more clearly than just poking around at your dataset with SQL. A clear use case we saw before was in the previous course looking at Cloud Dataprep. When you're in the transformer and you see those histograms at the top for the frequency of data values in that particular column, you get a very clear visual picture of a lot of that underlying data. What are the key trends? What are those anomalies that you see just by seeing those bar charts visually. Visualization can do a lot of things. Four quick things that we're going to highlight, you can spot those hidden trends. You can interact with a dataset, so you can tell a story that has multiple pieces. Say you start with a time series data and then you just click into one of the anomalous spikes or troughs in the data, and you drill down into those details very visually. It's a natural tool to let your audiences follow that cohesive story throughout the flow of your explanation. Of course, it can be very visually, aesthetically pleasing as well. Building those dashboards and conveying those insights in a very fast and effective manner. Lastly, a lot of your data is already in BigQuery, putting a visualization tool on top of it like Google Data Studio will naturally get all the performance benefits of having your data and your queries processed in BigQuery and then displayed and rendered quickly on the front end in a visualization tool. No lecture on data visualization will be complete without talking a little bit about visualization theory. As I mentioned before, visualization is both an art and a science. Here I'm going to get a little bit of the science of what our brains perceive when you look at a beautiful visualization. In this particular case, we have a stimulus where we have a cat. I immediately recognizes it and says, "Hey, I've seen that before, that's a cute little kitten," and your brain automatically says, "Hey, I've seen a thousand of these before," I'd immediately know without thinking that it's a cat. Now, when you actually get into machine learning, it's actually a little bit harder for machines to have that intuition. Intuition is extremely hard to build in for a computer. Whereas us humans we're built with a lot of that, what we call pre-attentive processing, where you can immediately recognize things. Now, what does it mean for your data visualizations? It means you can effectively cheat the brain by using common human intuition to not have the brain do a lot of work. Us humans have evolved to not do a lot of mental processing. We want to make those snap judgments very quickly and only tap into that really focused thought power when it's required. Let's take a look at an example together. Take a few seconds and count all the fives that are present. This is probably one of the hardest exercises that you're going to do as part of this specialization, you're looking at it. Did you get them all or are you still looking? Now, if you counted 16 fives in the the 10 seconds that I gave you, that's absolutely amazing. Must be on a speed reading level. For the majority of us myself included, it's very, very hard to pick out from this noise of numbers here, this very crowded visual what those fives are. Now, the meta point here is that this super focused processing that our brains have to do for the majority of us, we're reading from left to right and serially scanning every single row and counting out all those fives and keeping track of them. Now, there has to be an easy way to do that and naturally you might expect me to do something like this. Now, count the files. How about this time around, was it much easier? Now think about what your brain actually did to count those 16 and why it was much faster for you. On the surface you might say, well, you bolded the fives and that made them a lot easier to see. Well, that's absolutely true. Let's think of the theory behind that. In my mind, bolding enabled me to visually cluster those fives into small different chunks and quickly count them out and then pick them. The two things that stand out from a visualization theory perspective is when you contrast certain elements, you highlight the focus and it allows us to treat all the other numbers that aren't fives as background noise that we can safely ignore. In essence, it saves your brain's time by applying a prioritization saying, this is what's important, this is what you should focus on and immediately our brains can jump at the task and say, bam, "These are all the elements that I actually care about and I may ignore 80 percent of the rest of this visual here." We can take this yet a step further and continuously add more what we call visual encoding on these particular elements that we want to add focus. Here we introduce an element of color where you can further highlight the elements that you want folks to focus on and it gives it greater attention to what you're calling out as those key points. Now, there's a variety of methods that you can use to "cheat" the brain and then really tap into that fast processing time, that instant decision-making that we as humans have developed over the years. Now, you can mess around with things like the orientation, the shape and the skew, the length of a certain attribute, the size of the mark on the page, maybe some elements are curved and some aren't. You're adding things like a box around it, changing the intensity or the hue, moving around in the positioning or even adding things like emotion element to it as well. All of these will help the brain focus in on what's important. Then leveraging these and potentially in combination in concert with each other, will really help make your visualization stand out and ultimately convey that message very quickly without your audience having to stare at your screen and really focus hard.
