The book every Data Analyst should read

what up dead nerds i'm luke a data analyst and my channel is all about tech and skills for data science and let's talk about this book right here that i think every data nerd should read in order to level up their skills in communicating with data and judging by a recent youtube poll that i conducted a lot of y'all are not a fan of communicating with data so let's try to change that today for this i'm not going to be providing an in-depth review instead what i'm going to be doing is providing my insights from this specifically i'm going to be going over my five-step approach that i use in building any visualization in order to more effectively communicate with my audience so let's jump in my computer and look at a quick example of how i implement these skills that i learned from this book so say you're curious to find out what is the first skill you should be learning as a data analyst and judging by the subscribers of this channel you probably are so i could give you this visualization that shows the relative probabilities of a certain skill appearing in a linkedin job posting looking at it you can probably tell what the top skill is but can you tell what the second third or even fourth is let's add some data labels to clear the ambiguity this is better but i'd argue that it still takes too much time to interpret this graph and also what message are we trying to convey now what happens if we show this data but with a different graph this does a better job of not only ranking the values but also showing how they rank comparatively to each other this graph still doesn't answer your initial question of what skill should you learn first as a data analyst now with this final visualization we can thus more quickly get the reader to understand what insight we are trying to convey as a data analyst the more quickly and easily i'm able to convey a message or a data insight the more likely i'm going to be able to influence that audience member in how to make a decision so jumping into storytelling with data by colin affleck one interesting note on the author she used to work at google and she used to actually conduct seminars with employees of google on how to make better visualizations and those insights that she's learned over the years she built it and put it into this book full disclosure this video is not sponsored by cole or anybody associated with the book nor do they have any say in the making of this video this video does have a sponsor though and that is morning brew and now that i think of it we need to pay some bills so enjoy the sketch hey good morning luke oh hey what's up dad a nerd luke you know i don't like when you're yeah i know you don't like when i call you that hey what are you reading oh just getting my morning news from the journal how do you maintain the attention span to read that stuff um coffee dude remember the other day when i dropped those stats on the stock market and showed that funny meme that boss got a kick out of yeah where did you get those stats by the way well actually i got that from an article that i read on the morning brew that morning wait a new site that has memes well it's more of a free newsletter that you can read in less than five minutes and it's delivered straight to your inbox first thing in the morning free i pay ten dollars a month for this bologna sandwich yeah it's perfect for us data nerds because everything's backed up with data it even has a daily section that is dedicated to interesting statistics nice so those that are interested can use the link in the video description to sign up um what are you doing and i also just added a link in the comment section as well dude i've been using morning brew for forever now and i was really excited to share this and you totally just stole this moment from me thank you morning brew for sponsoring this video dude what [Music] and this is probably my last ever sponsored video all right getting back into my five-step approach in communicating with data the first step is understanding the importance of context and for this we need to ask three questions who you are communicating to what you want your audience to do and how can data help make that point so from that first example i gave a building a visualization let's compare that first graph that i made to that last graph that i made one quick note is that i am doing all these visualizations in power bi but you can really do this any graphing tools such as excel tableau or even python so for the who aspect you really need to think about who your target audience is when building a visualization for this we are trying to target aspiring data analysts so that is our target audience the who now moving into the what what do you want your audience to do for this they have the question of what skill do or should i learn first as a data analyst so for this we want to communicate what skills you should be learning as a data analyst and finally the how how can you use data to convey this point in this instance we're using data from linkedin that has relative probabilities of a skill appearing in a job posting for data analyst so this data is great at showing what skills are necessary of a data analyst so i would argue that that first and most definitely that last visualization that i made both of these achieve that importance of context or that step one of actually building your visualization and communicating your point but that one on the left needs to be refined further so let's move to step two for this we need to look into choosing an effective visual so why was that pie chart such an ineffective visual well if you break apart a pie chart and you start getting into comparing the relative sizes i feel a pie chart is really difficult to convey this point now i'm not saying pie charts are intrinsically bad there are a time and place for them especially if you have something like two attributes a pie chart may actually be pretty good for this however i would argue in this case maybe a doughnut chart along with some text visual would be better at conveying the point that you want to get across so what visuals should you be using for this in cole's book she outlines 12 different visualizations that are pretty routinely used i would take this a step further and say that for about eighty percent of the visuals that i use i'm either using a bar chart or a line chart and i think you should master these and i know what you're saying luke everybody uses bar and line charts and to that i say exactly these two type of visuals are frequently used because they're so quick at communicating insights additionally most people know how to read these visuals and so because of that you don't have to spend time teaching them how to interpret that visual for example let's say you're trying to determine what the best business intelligence software is and i present you with this visual here that shows the google trends results for both power bi and tableau over the past five years in a ribbon chart and yes the visualization does look pretty cool but are you actually getting your point across do they know the insight or what you want to convey in this i would say that most people are probably confused by this visual and you'd have to take the time to explain how this visual works instead i would argue that something like a line chart would be a much better visual in this case specifically will guide the reader to the results way quicker for this visualization you can more easily see their relative values compared to each other and also how they're trending over time and which one's on top and which one's on bottom if you need help selecting a visualization i'll include a link in the description below of a resource that i use called visual vocabulary where it's actually a tableau dashboard you can go in and select different visualizations based on your need all right moving into that third step and that is that you need to remove clutter so as great as excel is in its number crunching ability i'd argue that it's one of the biggest offenders in providing visualizations that are highly distracting take for example this visual that is showing counts of top skills for data scientist job postings on linkedin first the bar chart is fine but there's so many distracting aspects of this one of it being in 3d two with that blue background and then also three look how many skills they have you can't even read this for this let's say our audience is aspiring data scientist and they're trying to learn what the top programming language is to learn so let's remove all that clutter and just provide the basic insights from this and i'm only going to focus on the top 10 skills from here we can even further refine it to remove that legend because we don't really need it in this case so all these steps are working to remove clutter but you don't want to remove too much every axis should still have some sort of label in order to communicate what is going on with the data so we've completed that third step now of removing all that clutter we now need to get into the fourth aspect of focusing your audience's attention and the two ways that i like to focus my audience's attention is by size and color so let's remove all the color first and make it something more neutral and remember our goal is to communicate what is the top programming language to learn so let's use color to highlight these better now i've highlighted this in a way in order for the viewer's eyes to go to that first skill of python and then from there hopefully to that second and third skill of sql and r for coloring i typically like to leave everything in some sort of black or gray scale and then when i want to highlight something i'll use blue or shades of blue in order to get the viewer's attention now blue is my favorite color i also feel it's neutral so i like that color you feel free to use whatever color you want to use next up is size and for this i want to highlight the skills more than i want to highlight the counts so i'll adjust the formatting of the axises in order to make these differences now size can be used in the actual visual itself such as the size and a scatter plot or even like the case of our line chart that we did previously how bold of a line you want to make this all leads into the fifth and final step that i take and that is by adding text and titles in order to more easily convey my point this can be as simple as a title such as in the case of pinpointing what skill a data analyst should learn first for this i'm using that title above it and i'm also using that coloring in order to hint towards what they should be looking at or even in that similar case for the top programming languages for a data scientist we can use a title that also highlights where the reader should be interpreting and looking next now don't think that you're restricted to titles only you can also use text and other shapes to hint towards what insights you want to make from a graph for the tableau versus power bi visual it may be necessary sometimes to add context to what is happening so in this case i just tell what is directly happening over time all right so that's my five step approach i take when i'm communicating data insights to my colleagues or an audience i've found as a data analyst this has been super helpful for me in building effective visuals in order to convey the point that i'm trying to get across so all these insights were taken from this book of storytelling with data and this book itself goes into a lot more detail and reasoning behind a lot of these different decisions that i make so if you want to learn more about that sure check out the book with that as always if you got value out of this video smash that like button with that hey sorry about getting on you earlier about stealing my sponsor oh it's all right hey what are you doing um just a last-minute presentation build where'd you get those stats and where'd you get that meme dude what

Channel: Luke Barousse

Views: 200,433

Keywords: data viz by luke, business intelligence, data science, bi, computer science, data nerd, data analyst, data scientist, how to, data project, data analytics, portfolio project, sql, excel, python, power bi, tableau, data engineer, storytelling with data, cole knaflic, cole nussbaumer knaflic, data visualization, data viz, graphs, charts

Id: 09JnFEdZe2A

Length: 10min 34sec (634 seconds)

Published: Tue Feb 08 2022