Data Storytelling with Google Looker Studio E-Book

Sireesha Pulipati

• Herausgeber: Packt Publishing
• Kategorie: Wissenschaft und neue Technologien
• Sprache: Englisch

Seitenzahl: 528

Veröffentlichungsjahr: 2022

Beschreibung

Presenting data visually makes it easier for organizations and individuals to interpret and analyze information. Looker Studio is an easy-to-use, collaborative tool that enables you to transform your data into engaging visualizations. This allows you to build and share dashboards that help monitor key performance indicators, identify patterns, and generate insights to ultimately drive decisions and actions.
Data Storytelling with Looker Studio begins by laying out the foundational design principles and guidelines that are essential to creating accurate, effective, and compelling data visualizations. Next, you’ll delve into features and capabilities of Looker Studio – from basic to advanced – and explore their application with examples. The subsequent chapters walk you through building dashboards with a structured three-stage process called the 3D approach using real-world examples that’ll help you understand the various design and implementation considerations. This approach involves determining the objectives and needs of the dashboard, designing its key components and layout, and developing each element of the dashboard.
By the end of this book, you will have a solid understanding of the storytelling approach and be able to create data stories of your own using Looker Studio.

Leseprobe

Data Storytelling with Google Looker Studio

A hands-on guide to using Looker Studio for building compelling and effective dashboards

Sireesha Pulipati

BIRMINGHAM—MUMBAI

Data Storytelling with Google Looker Studio

Copyright © 2022 Packt Publishing

All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, without the prior written permission of the publisher, except in the case of brief quotations embedded in critical articles or reviews.

Every effort has been made in the preparation of this book to ensure the accuracy of the information presented. However, the information contained in this book is sold without warranty, either express or implied. Neither the author, nor Packt Publishing or its dealers and distributors, will be held liable for any damages caused or alleged to have been caused directly or indirectly by this book.

Packt Publishing has endeavored to provide trademark information about all of the companies and products mentioned in this book by the appropriate use of capitals. However, Packt Publishing cannot guarantee the accuracy of this information.

Publishing Product Manager: Heramb Bhavsar

Senior Editor: Nazia Shaikh

Content Development Editor: Sean Lobo

Technical Editor: Rahul Limbachiya

Copy Editor: Safis Editing

Project Coordinator: Farheen Fathima

Proofreader: Safis Editing

Indexer: Tejal Daruwale Soni

Production Designer: Ponraj Dhandapani

Marketing Coordinators: Priyanka Mhatre and Nivedita Singh

First published: October 2022

Production reference: 2311022

Published by Packt Publishing Ltd.

Livery Place

35 Livery Street

Birmingham

B3 2PB, UK.

ISBN 978-1-80056-876-1

www.packt.com

To my father, Seshagiri Rao Pulipati, who is my north star and inspiration. To my husband, Uma Shanker Avvari, for being the anchor of my life and of everything I do and achieve.

– Sireesha Pulipati

Foreword

Some years ago, I was standing on a career precipice. I was in the middle of my first ever product launch, which was being crowd funded on KickStarter. The product was a low tech boardgame approach to creating wireframes for business dashboards; it was called the Dashboard Wireframe Kit. This was a completely new environment to me, having left the world of Big 4 analytics consulting and plunged into the foray of being an independent consultant with a heavy specialization in data visualization and driving the adoption of dashboards in enterprise.

One of the very early supporters of what I was doing was the author of this book. She was a vocal advocate for storytelling with data, though what really stood out was her practical perspective on how to go about it. She had a perspective that could only be borne out of years of working in the field and actually having to figure out what works, and what was best left as theory, from the many books out there on the topic.

I’ve read these books, and many of them present excellent theory. Others present thoughtful technical approaches to using certain specific tools. Rarely do you come across a work that combines the two. Yet, that is what the author has achieved here – a highly valuable distillation of data visualization best practices, with a targeted technical application to Google Looker Studio. Put down whatever else you are reading, because this book compresses a vast amount of knowledge into a potent knowledgebase that will make you a proficient in both visualization, and in applying that expertise directly in the tool. Enjoy!

Nicholas Kelly Author | Delivering Data Analytics

Contributors

About the author

Sireesha Pulipati is an experienced data analytics and data management professional. She has spent the last decade building and managing data platforms and solutions, and she is passionate about enabling users to leverage data to solve business problems. Sireesha holds a master’s degree in business administration and a bachelor’s degree in electrical engineering. Her work history spans multiple industries – healthcare, media, travel and hospitality, high-tech, and more. She is currently at Google as a technical lead, helping with the business intelligence and analytics strategy for the teams building and supporting the Knowledge Graph. Outside of work, Sireesha enjoys hiking and reading books. She currently resides in the San Francisco Bay Area.

About the reviewers

Ganesh is a self-motivated IT professional with extensive experience in leading and managing information technology projects in companies such as Oracle and Standard Chartered Bank. He has completed a postgraduate degree with computer science as his major at Bharathidasan University, and executive education in finance from the Indian Institute of Management, Kozhikode. As a passionate data visualization expert, he has designed beautiful eye-catching dashboards in both Google Looker Studio and Zoho Analytics for various clients. He also specializes in the development of applications using the Zoho suite of products.

I would like to thank Packt Publishing for giving me the opportunity to review this book. I thank my family, friends, and all the well-wishers for supporting me all these years.

Ralph Spandl moved to Montreal after finishing his studies at the HdK (University of Fine Arts) in Berlin, where he founded r42, which slowly became a web agency with a nice portfolio of companies, working in the financial, industrial, and leisure sectors and a few non-profit organizations. For years, Ralph had a soft spot for data visualization, which he planned to center his professional career around. Since 2018 or thereabouts, he has been doing just that, designing and coding data visualizations, and he is now working with Supermetrics on the implementation of the first commercial Looker Studio chart library, a collection of data visualizations that allow you to push the limits of Looker Studio.

Table of Contents

Preface

Part 1 – Data Storytelling Concepts

1

Introduction to Data Storytelling

Understanding data storytelling

Structuring a data story

Presenting data stories

Data storytelling at work

Essential skills for data storytelling

Building data stories – an approach

Determine

Design

Develop

Walking through an example

Summary

2

Principles of Data Visualization

Understanding foundational design principles

Simplicity of design

Organizing the layout

Accuracy of information presented

Reviewing Gestalt principles of visual perception

Proximity

Similarity

Continuity

Closure

Enclosure

Connectedness

Prägnanz

Figure and ground

Focal point

Using color wisely

Use fewer distinct colors

Choose an appropriate color palette and scheme

Use color consistently across the dashboard

Consider inclusive color schemes

Summary

Further reading

3

Visualizing Data Effectively

Choosing the right visuals

Scorecards

Pie and donut charts

Bar charts

Line charts

Combo charts

Scatterplot

Tables

Heatmap (matrix)

Treemap

Geographical maps

Others

Avoiding common pitfalls

Overloading a dashboard

Designing a poor or incohesive layout

Not emphasizing key information and a message

Using color excessively or inappropriately

Using dual axes in charts without caution

Inappropriate manipulation of axes

Using inappropriate or complex chart types

Summary

Further reading

Part 2 – Looker Studio Features and Capabilities

4

Google Looker Studio Overview

Technical requirements

Getting started with Google Looker Studio

How it works

Working with data sources

Creating a data source

Managing data freshness

Controlling data access

Editing a data source schema

Other common data source operations

Working with reports

Creating a report

Creating a report from a template

Publishing a report

Sharing a report

Other common report operations

Working with Explorer

Creating an Explorer

Exporting from Explorer

Using Explorer in an analyst workflow

Leveraging Looker Studio Gallery

Getting help with Looker Studio

Building your first Looker Studio report – creating the data source

Summary

5

Looker Studio Report Designer

Technical requirements

Report Designer overview

Adding charts to the canvas

Adding additional data sources

Adding and managing pages

Choosing a report theme and layout

Defining Report Settings

Working with data for charts

Adding dimensions

Adding metrics

Sorting data in the charts

Implementing filters

Understanding editor filters

Adding an editor filter

Interactive filter controls

Adding design components

Embedding external content

Styling report components

Background and Border

Text styles

Common chart style properties

Configuring style properties in report themes

Building your first Looker Studio report – creating a report from the data source

Summary

6

Looker Studio Built-In Charts

Technical requirements

Charts in Looker Studio – an overview

Configuring tables and pivot tables

Table with numbers

Table with bars

Table with drill down

Pivot tables

Configuring bar charts

Columnar bar chart

Horizontal bar chart

Clustered bar chart

Stacked bar chart

Configuring time series, line, and area charts

Line chart

Time series chart

Area chart

Configuring scatter charts

Configuring pie and donut charts

Configuring geographical charts

Geo chart

Google Maps chart

Configuring scorecards

Configuring other chart types

Treemap

Bullet chart

Gauge chart

Building your first Looker Studio report – adding charts

Summary

7

Looker Studio Features, Beyond Basics

Technical requirements

Leveraging calculated fields

Organizing dimension values into custom groups

Manipulating text with regular expressions

Using MAX and MIN across multiple fields or expressions

Displaying images and hyperlinks

Using parameters

Parameters and calculated fields

Parameters and connectors

Blending data

Blending disparate data sources

Blending charts

Reaggregating metrics using blending

Adding community visualizations

Creating report templates

Optimizing reports for performance

Optimizing data sources

Optimizing reports

Underlying dataset performance

Summary

Part 3 – Building Data Stories with Looker Studio

8

Employee Turnover Analysis

Technical requirements

Describing the example scenario

Building the report - Stage 1: Determine

Building the report - Stage 2: Design

Defining the metrics

Choosing the visualization types

Considering the filters and their interactions

Designing the layout

Building the report - Stage 3: Develop

Setting up the data source

Creating the report

Summary

9

Mortgage Complaints Analysis

Technical requirements

Describing the example scenario

Introducing BigQuery

Getting started with BigQuery

Getting data into BigQuery

Analyzing data in BigQuery

Building the dashboard- Stage 1: Determine

Building the dashboard- Stage 2: Design

Choosing visualization types

Considering filters and interactions

Designing the layout

Building the dashboard- Stage 3: Develop

Setting up the data source

Creating a report

Summary

10

Customer Churn Analysis

Technical requirements

Describing the example scenario

Building the dashboard- Stage 1: Determine

Building the dashboard- Stage 2: Design

Defining the metrics

Choosing visualization types and filters

Designing the layout

Building the dashboard- Stage 3: Develop

Setting up the data source

Creating a report

Summary

11

Monitoring Report Usage

Technical requirements

Usage monitoring overview

Google Analytics primer

Understanding GA reports

Monitoring Looker Studio report usage with GA4

Setting up GA4 for Looker Studio report monitoring

Creating a custom report in GA4

Visualizing in Looker Studio

Exporting GA4 data to BigQuery

Summary

Index

Other Books You May Enjoy

Preface

Organizations and individuals are increasingly relying on data to make important decisions. Presenting data visually makes it easier to interpret and analyze. Google Looker Studio is an easy-to-use and collaborative tool that helps you explore your data and transform it into beautiful visualizations. With Looker Studio, you can build and share dashboards that help monitor key performance indicators, identify patterns, and generate insights that ultimately drive decisions and actions.

The goals of this book are threefold: first, provide foundational know-how on basic design and visualization principles, second, offer a practical and demystified guide on using Looker Studio for visualizing data, and third, give a walk-through of the structured dashboard building process and the various deliberations involved in it. Data Storytelling with Google Looker Studio begins with laying out the foundational design principles and guidelines that are essential to creating accurate, effective, and compelling data visualizations. We then delve into the features and capabilities of Looker Studio – from the basic to the advanced – and showcase their application with examples. The book then takes you through the process of building dashboards with a structured three-stage process called the 3-D approach using real-world examples. The approach involves determining the objectives and needs of the dashboard, designing its key components and layout, and developing each element of the dashboard. These examples take you through the thought process of various design and implementation considerations.

Reports and dashboards are two forms of presenting data visuals together. They fundamentally serve different purposes and differ in terms of level of detail, interactivity, breadth and so on. However, for all practical purposes of this book, the distinction between the two doesn't matter too much. Hence, I use the terms report and dashboard interchangeably through much of this book. In cases where the distinction makes a difference to the topic discussed, I call that out specifically.

All through the writing of this book, right up to its publication, the tool we used was called “Data Studio.” Google announced the rebranding of the tool as “Looker Studio” on October 11, 2022, which reflected on the tool itself as well as the associated documentation and support pages almost instantaneously, or so it seemed.

Google acquired Looker, the new-age enterprise Business Intelligence (BI) and data analytics platform, in 2019. With its logical semantic layer, in-database architecture, API and developer-friendly capabilities, Looker provides a powerful platform to meet enterprise business intelligence needs. Looker became part of the Google Cloud offerings, and it complemented the existing free Data Studio tool. Together, the two business intelligence tools provided flexibility and choice to the users.

The rebranding is part of a strategy to consolidate all Google Cloud's business intelligence services under the Looker brand. The Data Studio tool itself remains the same, and there is no change in its capabilities and features as a result of this. From a UI standpoint, only the logo is changed. Also, the product is still free. Google has introduced a new premium tier to Looker Studio, called Looker Studio Pro, with additional capabilities and support that cater to enterprise teams.

This book is only limited to the free Looker Studio (formerly, Data Studio) tool and does not touch upon any enterprise capabilities of the Pro version of Looker Studio or the Looker platform. While an attempt is made to use the new name - Looker Studio - as much as possible throughout the book, screenshots and images mostly reflect the old name and logo.

A big part of this strategic move by Google is the strong integration between Looker Studio and the Looker Platform. Looker enables you to create semantic models of your data by defining relationships between data sets, creating metrics, and encapsulating business logic. With the new Looker connector, you can connect to your Looker models from Looker Studio and visualize the data, without you needing to build the relationships, creating metrics, or formatting fields within Looker Studio. While the connector is free, you need a valid license and appropriate permissions to the Looker Platform to connect. Looker is in turn very deeply integrated with BigQuery. Looker by itself does not store any data. It connects to the data stored in BigQuery, and provides a logical layer on top of it to meet the data exploration, analytical, and reporting needs of the users. It thus leverages the powerful analytical capabilities of BigQuery. The Looker platform has its own visualization layer, which is complementary to Looker Studio. As a BI enthusiast, I'm very excited about this direction that Google has taken with its BI portfolio and I will closely follow its evolution - you should too.

Who this book is for

If you are a beginner or an aspiring data analyst looking to understand the core concepts of data visualization and you want to use Google Looker Studio for creating effective dashboards, this book is for you. No specific prior knowledge is required to benefit from this book.

If you are a more experienced data analyst or business intelligence developer, you will find this book useful as a detailed guide to using Looker Studio as well as a refresher of the core dashboarding concepts.

If you are a business professional looking to build reports and run analyses on your own, this book empowers you with the knowledge and skills you need to visualize data effectively using the simple and easy-to-use tool Looker Studio.

What this book covers

Chapter 1, Introduction to Data Storytelling, introduces the concept of data storytelling, its format, and its manifestation in dashboards and reports.

Chapter 2, Principles of Data Visualization, covers foundational principles and guidelines that enable the creation of effective and compelling data visualizations.

Chapter 3, Visualizing Looker Effectively, describes some common chart types and their applications along with pitfalls to avoid.

Chapter 4, Google Looker Studio Overview, gets you started with Looker Studio and describes how to work with and manage key entities such as data sources, reports, and explorerss.

Chapter 5, Looker Studio Report Designer, examines key report designer options, settings, and elements such as report theme, pages, filter controls, styling, and more that help design reports in Looker Studio.

Chapter 6, Looker Studio Built-In Charts, reviews the built-in charts provided by Looker Studio and their configurations.

Chapter 7, Looker Studio Features, Beyond Basics, covers advanced features such as calculated fields, parameters, blending, report templates, community visualizations, and report optimization.

Chapter 8, Employee Turnover Analysis, walks you through building a detailed report analyzing employee turnover for a fictious company using the 3-D approach: Determine, Design, and Develop.

Chapter 9, Mortgage Complaints Analysis, walks you through building a dashboard for monitoring mortgage-related complaints received by the Consumer Financial Protection Bureau (CFPB), a US agency, using the 3-D approach.

Chapter 10, Customer Churn Analysis, walks you through building a dashboard to analyze the customer churn phenomenon for a broadband service company using the 3-D approach.

Chapter 11, Monitoring Looker Studio Report Usage, describes how to track and monitor usage of Looker Studio reports using Google Analytics.

To get the most out of this book

Looker Studio is a web-based tool. You need a Google account and a supported browser to follow along and benefit from the book. Basic SQL knowledge will help you explore a few topics, but is not mandatory. Access to a Google Cloud Platform account, either a free trial or paid, is nice to have and will help you visualize data from BigQuery public datasets. You can leverage the free BigQuery sandbox for this purpose as well.

Software/hardware covered in the book

Operating system requirements

Looker Studio (web-based)

NA

Google Cloud Platform subscription (free trial or paid) or BigQuery sandbox (free)

NA

Google Analytics (web-based)

NA

Google Cloud Platform is used to demonstrate visualizing data from BigQuery, Google’s petabyte-scale cloud data warehouse. It is leveraged only in a couple of chapters in the book. No prior knowledge of BigQuery is expected. The details of how to get started with it and connect to it from Looker Studio are included in Chapter 9, Mortgage Complaints Analysis. Google Analytics is a free Google tool and is used to monitor the reports of Looker Studio in Chapter 11, Monitoring Looker Studio Report Usage.

If you are using the digital version of this book, we advise you to type the code yourself or access the code from the book’s GitHub repository (a link is available in the next section). Doing so will help you avoid any potential errors related to the copying and pasting of code.

Download the example code files

You can download the example code files for this book from GitHub at https://github.com/PacktPublishing/Data-Storytelling-with-Google-Data-Studio. If there’s an update to the code, it will be updated in the GitHub repository.

We also have other code bundles from our rich catalog of books and videos available at https://github.com/PacktPublishing/. Check them out!

Download the color images

We also provide a PDF file that has color images of the screenshots and diagrams used in this book. You can download it here: https://packt.link/5u31Q.

Conventions used

There are a number of text conventions used throughout this book.

Code in text: Indicates code words in text, database table names, folder names, filenames, file extensions, pathnames, dummy URLs, user input, and Twitter handles. Here is an example: “BigQuery provides this information as part of the census_bureau_acs public dataset.”

A block of code is set as follows:

CREATE TABLE `datastudio-343704.data_viz.baseball_schedule` AS SELECT * FROM `bigquery-public-data.baseball.schedules

Bold: Indicates a new term, an important word, or words that you see onscreen. For instance, words in menus or dialog boxes appear in bold. Here is an example: “The SETUP tab is where you choose the appropriate data source and add different fields – dimensions and metrics that make up the chart.”

Tips or important notes

Appear like this.

Get in touch

Feedback from our readers is always welcome.

General feedback: If you have questions about any aspect of this book, email us at [email protected] and mention the book title in the subject of your message.

Errata: Although we have taken every care to ensure the accuracy of our content, mistakes do happen. If you have found a mistake in this book, we would be grateful if you would report this to us. Please visit www.packtpub.com/support/errata and fill in the form.

Piracy: If you come across any illegal copies of our works in any form on the internet, we would be grateful if you would provide us with the location address or website name. Please contact us at [email protected] with a link to the material.

If you are interested in becoming an author: If there is a topic that you have expertise in and you are interested in either writing or contributing to a book, please visit authors.packtpub.com.

Share Your Thoughts

Once you’ve read Data Storytelling with Google Data Studio, we’d love to hear your thoughts! Please click here to go straight to the Amazon review page for this book and share your feedback.

Your review is important to us and the tech community and will help us make sure we’re delivering excellent quality content.

Part 1 – Data Storytelling Concepts

This part introduces the concept of data storytelling, its importance, and its purpose. It presents the fundamental design principles that are key to building effective visualizations. It also looks at the basic dos and don’ts of data visualization and dashboard/report design. You will gain the foundational knowledge that’s needed before you begin your own data storytelling journey.

This part comprises the following chapters:

Chapter 1, Introduction to Data StorytellingChapter 2, Principles of Data VisualizationChapter 3, Visualizing Data Effectively

1

Introduction to Data Storytelling

Data storytelling makes the presentation of data compelling and persuasive. This is a book about learning how to use theLooker Studio tool to visualize data and build useful reports and dashboards. Before we learn how to build different visual data representations, it’s important to first understand the craft of storytelling with data and its utility. It will serve us well to take a step back and understand the big picture.

This introductory chapter sets the stage for understanding the meaning of data storytelling and its importance. We will go through the components of a good data story, understand how data storytelling manifests at work, and learn the key skills required to be a good storyteller. Finally, we will learn about an approach to building effective data stories. In this chapter, we will cover the following main topics:

Understanding data storytellingBuilding data stories – an approach

Understanding data storytelling

Simply put, data storytelling means telling a story using data. According to Brent Dykes, author of the bestselling book Effective Data Storytelling, data storytelling is the skillful amalgamation of data, narrative, and visuals.

Why tell a story? Stories are an integral part of our lives and are the most natural way we consume and retain information. Compared to straightforward facts and messages, stories are memorable, impressionable, relatable, and persuasive. Stories appeal to the humanness of the audience. Stories often help in communicating a complex concept or message more effectively. This is evident through the prevalence and effectiveness of parables, fables, and other forms of stories throughout human history.

Structuring a data story

Stories can be told in different ways and can have different purposes. Some have a moral or a lesson to teach, some report events, while others just entertain. In the simplest sense, a story has a beginning, a middle, and an end. This basic structure is referred to as Aristotle’s arc, known to have been introduced by the ancient Greek philosopher.

Figure 1.1 – Aristotle’s arc

The traditional narrative arc expands on this basic linear structure and involves building up tension through rising and falling action. German playwright Gustav Freytag’s five-step paradigm, commonly known as Freytag’s Pyramid, forms the basis of most modern-day stories.

Figure 1.2 – Traditional story structure – Freytag’s Pyramid

In this traditional form of narrative, the most important aspects of the story are revealed in the latter half of the story – climax through resolution. The traditional story approach when applied to data stories typically involves the following narrative flow:

Provide the contextState the problemHighlight the impact Share the key insight

The other major form of storytelling is news reporting. It follows an inverted pyramid format. The most important information is provided at the beginning, followed by the key supporting information, and finally, the least important details.

Figure 1.3 – News story structure

When a data story is created using the news story approach, it follows the following sequence of steps:

Share the key insightProvide the context and causesDetails through drill-throughs/drill-downs

A data story can use either of these approaches depending on the presentation format, the audience, and the objective of the narrator – to influence a decision, to inform, to trigger an action, and so on. If you are presenting data to an executive audience with little time to spare, you might want to follow the news story approach by starting with the key insight. This will help you get their attention and then you can delve into the details as needed.

On the other hand, when you are trying to present a complex or counter-intuitive insight, you might want to follow the more traditional approach of first setting the stage with the context and evidence, then laying out the problem, drawing attention to the impact of this problem, and finally closing with one or more solution recommendations.

The purpose of a data story could either be explanatory, where we explain a phenomenon, or be actionable, where we want to elicit an action or decision through actionable insights. While data can also be used to describe a situation, the descriptive nature doesn’t typically make a story by itself. Data on its own has no useful meaning. It needs to be gleaned for information and insights. And it is these insights that we are usually after.

Instances of powerful data stories that persuade action and influence decision making are ubiquitous. Consider a non-profit organization that is seeking donations for supporting cancer research. Sharing personal stories of those who suffered from cancer and those who benefitted from the research supported by the organization makes an incredible impact on the potential donors. Presenting a data story around the effectiveness of the organization, the amount of money raised, the membership growth, and the people served helps potential donors and volunteers to connect with the cause and persuade them to take action.

As another example, consider the owners of an online personalized gift store. Their sales are declining and they would like to understand what action they can take to remedy the situation. By analyzing sales and customer feedback data, they learned that over 30% of customers in the last 6 months have experienced shipping delays and damage. These customers left poor reviews on the e-commerce site and prominent social media forums, resulting in poor sales. Based on this insight, the owners decide to replace the shipping carrier. This caused the negative customer feedback to dwindle immediately and the store saw more positive feedback and sales over time.

Data storytelling is not the same thing as data visualization. While data visualization refers to the effective representation of data through graphics and visuals, data storytelling goes beyond just data visualization. Visuals are critical but are only one component of data storytelling. Data storytelling embeds data visuals in a narrative and presents a cohesive picture.

Note

Much of this book is focused on building great visualizations with Looker Studio. We will touch upon the narrative aspects of data storytelling where applicable.

When building a data story, always start with understanding your data and identifying the key insight or phenomenon you would like to share. Then create a narrative that you would like to take your users or audience through. Follow that with sketching the scenes and designing the storyboard. Only then work on building the visuals and presentation.

Further reading

For a deeper understanding of the psychology of storytelling and various aspects of data stories, read Brent Dykes’ book Effective Data Storytelling.

Presenting data stories

Data stories can be presented either to a live audience for direct consumption or to an offline audience to consume the content indirectly. Data stories can take many forms – documents, PowerPoint presentations, videos, websites, dynamic dashboards, reports, and more. When you present directly to the audience, you are in full control of the narrative. You determine what the audience sees at any point in time. You can carefully walk them through your various story scenes in sequence, building up the necessary tension and anticipation.

In this mode, you can also employ various visual aids and tools – images, animations, video, text, charts, and more – to facilitate the narrative and make the presentation more compelling. The audience passively consumes the information you are providing. Nevertheless, it can be presented in quite an engaging way. Perhaps the best example of data storytelling comes from the Swedish physician and public speaker Hans Rosling’s iconic narration of the story of the world using augmented reality animation. You can watch it on YouTube at https://www.youtube.com/watch?v=jbkSRLYSojo.

On the other hand, when the intent is to present to an audience or users who will consume the content later, the format can be static, such as in published reports or articles, or can be interactive, such as on dashboards, websites, and so on, allowing the audience to interact with the content and explore. In offline consumption mode, you need to be really cautious about driving the desired user behavior so that all users can interpret and understand the key insight or phenomenon consistently and with little ambiguity.

When the data is static and doesn’t change over time, as the narrator or storyteller, you know exactly the insight or the message that needs to be conveyed and how best to present it. A good example of using text and visuals to narrate a story about child mortality data can be found at https://vizhub.healthdata.org/child-mortality.

With the help of simple animations, powerful visuals, and supporting text, the narrator has built a compelling narrative to highlight the problem of child mortality and how uncovering inequalities in child survival accelerates progress toward achieving the Global Sustainable Development Goals for child mortality (https://data.unicef.org/topic/child-survival/child-survival-sdgs/).

If data is not static and is updated on a regular basis, you are not in control of what the narrative will be because the insights may change over time. This is usually the case with business reports and dashboards. For this reason, it is static data that is the most amenable to creating data stories. However, you will still benefit from the data storytelling approach to develop these dashboards to create as cohesive a narrative as possible. Business dashboards are usually good at serving descriptive and explanatory purposes. However, you can also strive to help users identify any insights easily through intelligent design.

Figure 1.4 – Source: Looker Studio Report Gallery

This screenshot is an example of a well-designed dashboard. It is built using Google Analytics data and is available for use as a template from the Looker Studio Report Gallery.

Report versus dashboard

While I often use the terms report and dashboard together or interchangeably in this book, they actually represent two distinct forms of presenting data and generally serve different purposes. However, the concepts I discuss in this book apply to both constructs.

The following table lists some major differences between reports and dashboards:

Table 1.1 – Differences between a report and a dashboard

Note

In Looker Studio, there is only a Report object, using which you can build either detailed reports or high-level dashboards based on the audience and the purpose at hand.

In the next section, we will discuss the role data storytelling plays in organizations.

Data storytelling at work

Being data-driven is the hallmark of all successful organizations. Thomas Davenport, the world-renowned thought leader, published his groundbreaking work Competing on Analytics in 2006. Ever since then, companies across industries have embraced analytics and embarked on the analytics journey to truly differentiate themselves from the competition. They achieved this by linking analytics with decision making. According to Davenport, there are five stages along the path to competing successfully in analytics:

Analytically impairedLocalized analyticsAnalytical aspirationsAnalytical companiesAnalytical competitors

Irrespective of where an organization is on this journey, there are several avenues to leverage data and analytics to generate insights, persuade people, or influence and aid decision making. Analytics can help understand what is happening across different business functions viz. finance, customers, marketing, sales, and so on, and why. They can help determine the effectiveness of projects and programs across the organization to enable better decisions regarding resources. They can provide insights into optimizing people processes and allow for a data-driven approach to address human capital issues. Analytics can also be useful on a personal level – to influence your leadership or peers using data. Some examples of these different types of analytics are listed as follows:

Business analytics: Financial performance, customer engagement, product analytics, campaign effectiveness, sales pipelines, and so onProject/program analytics: Resource allocation, avoiding sunk cost fallacy, alignment with key results and objectives, and so onPeople analytics: Employee attrition, absenteeism, diversity, equity, and inclusion (DEI) compliance, recruitment and hiring efficiency, employee engagement, learning and development effectiveness, and so onManage up/down: Bring data and analytics that showcase your impact to the performance review conversations with your manager, tell a story backed by data to pitch your idea to peers, and so on

A critical capability that enables organizations to progress through the five stages of competitive data analytics and become truly data-driven is data literacy. Data literacy is the ability to read, understand, communicate, and work with data effectively. Thinking critically about the data and the insights it generates is also a core aspect of data literacy. For an organization to achieve a competitive advantage through data and analytics, its employees need to have high levels of data literacy and it has to be more prevalent across the organization. The main components of data literacy are as follows:

Understand: Know what the data represents and the purpose it serves; understand key metrics.Read: Interpret graphical and visual representations of data correctly.Produce: Wrangle, clean, and prepare data appropriately to meet the analytical needs.Communicate: Convey the insights effectively through the use of various tools and techniques.Critical thinking: Be curious about the data and what it can tell us and think critically about the insights it generates.

Essential skills for data storytelling

Data storytelling specifically addresses the Communicate aspect of data literacy. But it also embodies the other components as you cannot generate insights and communicate effectively without first understanding the data, working with the data, and thinking critically about the data. So, the key skills required to be a good data storyteller are as follows:

Get and prepare the dataIdentify the right data to address the problem at hand and obtain it. You might have to work with other teams to get access to that data. Or in the case of external data, you might have to make an API call, scrape the web, or download the data yourself.Clean the data, format it, and munge it appropriately (reshaping, aggregating, filtering, and so on).Analyze the data and generate insightsExplore the data and understand the trends and relationships among different attributes. Create new metrics and apply analytical methods to glean insights.Visualize the dataVisualize the data in appropriate charts. You’ll need to have the technical skill to use a specific tool – be it Google Sheets, Google Slides, MS Excel, MS PowerPoint, Looker Studio, Looker, Tableau, and so on.Build a narrativeChoose an appropriate narrative structure and design the storyboard.

Data storytelling is an essential skill that allows for effective communication through data. Simplicity can be thought of as the mantra of data storytelling so that anyone and everyone can understand the message and insight clearly and unambiguously. However, certain complex problems and phenomena might require a little more advanced representation and nuanced interpretation. A higher level of data literacy among stakeholders ensures that such data stories drive quality decision making.

At work, data storytelling usually manifests in the form of static slide decks, documents, and spreadsheets or dynamic dashboards and reports. The next section explains an approach that we can follow to build effective dynamic dashboards and reports based on the tenets of data storytelling described in the first section of this chapter.

Building data stories – an approach

Business dashboards and reports pose unique challenges in crafting a story owing to the dynamic nature of the data they support. In this section, we are going to learn about the 3-D approach to designing and building effective dashboards. The approach consists of three main stages as shown in the following figure:

Figure 1.5 – Stages of the data storytelling approach

In the Determine stage, you need to work on determining the audience for the dashboard, the key business questions they need answering, and the data required to answer those questions. In the Design stage, you will work on the narrative, select the right visuals, and identify the required interactive controls. In the Develop stage, you create the visuals and implement the right colors and visual cues for a seamless user experience. In this final stage, you will choose the appropriate delivery method and accordingly deploy and share the dashboard or the report.

Let’s go through each of these stages in the rest of this section.

Determine

When embarking on the journey of building a dashboard, start by determining the following:

Users: Who are the primary users of the dashboard? Interview key target users to capture the key business objectives, challenges they face, and questions they need answering in their day-to-day work. Consolidate these requirements and define personas. Among other relevant information, a persona captures the following:Name and roleResponsibilities/objectivesNeeds and expectationsChallenges

Ideally, a dashboard should be built to fit only a single persona. Trying to meet the needs of multiple personas with a single dashboard is a recipe for disaster, which could leave all users less than satisfied. On the other hand, a report can accommodate the needs of multiple personas and be as comprehensive and detailed as needed.

Business questions: Identify the key business questions from the identified persona(s) and define critical analytical user journeys. Have a clear understanding of the problem the dashboard is going to address and the supporting evidence required.Data: Find the data needed to address the business questions and provide additional context. Determine the appropriate level of data granularity needed. Make sure that the data can be accessed. If the right data is not available to provide complete and accurate answers to the previous business questions, adjust the purpose of the dashboard appropriately and set the right expectations with the end users.

Design

After determining the target users, well-defined business questions, and the data required to build the dashboard, the next step is to work on the design. Designing a dashboard involves the following aspects:

Design the right metrics: Break down the main question into key subcomponents and define the right metrics. Determine which metric types – absolute numbers, differences, percentages, ratios, and so on – are the most appropriate. For example, ratios or percentages to enable comparisons, simple aggregations to summarize data, and so on. Identify any data preparation and manipulation needs in order to build these metrics.Choose the right visualization type: Determine the right chart types to use with different metrics and highlight appropriate relationships and associations. In the next chapter, we’ll review some common chart types and general guidelines as to when each is appropriate.Design the narrative: Determine the flow of information and craft the story. Sketch the layout and organize various charts to fit the narrative. User experience studies have shown that most users consume content from top left to bottom right in a Z path. Accordingly, place the most important visuals at the top, and add supporting charts below them, creating a logical flow.Design interactivity: It is also in the design stage that you determine the interactive filters that you need to provide so that the users can make appropriate selections and view different slices of data. Provided the tool of choice offers the capability, you can also design cross-filtering capabilities between related charts to help users better understand the data. This is also an appropriate time to determine whether the users could benefit from implementing any alerts – when a metric raises above or falls below a threshold, a notification can be sent out to the users so that they are informed of the discrepancy or anomaly automatically and in a timely fashion. This enables users to take appropriate and timely action.

The outcome of this stage can be anything from a simple hand-sketch to a mockup implementation built with dummy data.

Develop

Once you have crafted the narrative, designed the layout, identified the right metrics, and chosen the appropriate visualizations, it’s time to develop the dashboard by implementing the design. You start with connecting to the data and building the metrics. You then develop each individual chart and arrange them according to the layout design. In this stage, you choose the right colors, themes, text, annotations, and other visual aspects that provide a desirable user experience. Implement alerts to trigger notifications to users as needed.

Once you finish developing the dashboard, you deliver it to end users by deploying it and providing access. Users can consume the dashboard either directly by accessing the tool on which it is built or subscriptions can be set up so that they can receive the latest snapshot of the dashboard to their email inboxes at a suitable frequency. As part of this stage, you will also make sure the dashboard is refreshed with the latest data at a pre-determined frequency, which is largely dependent on the freshness of the underlying data source.

Walking through an example

Let’s deepen our understanding of the 3-D data storytelling approach to building dashboards with an example. A UK-based online retail company sells unique all-occasion gifts to its largely wholesale customer base. It caters to customers across many countries. We are going to build a dashboard in Looker Studio based on its online sales data. The data is made available by UC Irvine in its Machine Learning Repository (http://archive.ics.uci.edu/ml/datasets/Online+Retail). The Online Retail dataset has been donated to UCI Machine Learning repository by Dr. Daqing Chen of London South Bank University.

Citation

Daqing Chen, Sai Liang Sain, and Kun Guo, Data mining for the online retail industry: A case study of RFM model-based customer segmentation using data mining, Journal of Database Marketing and Customer Strategy Management, Vol. 19, No. 3, pp. 197–208, 2012 (Published online before print: 27 August 2012. doi: 10.1057/dbm.2012.17).

To build this example dashboard, I've added a new column to this data. You can access the modified dataset at https://github.com/PacktPublishing/Data-Storytelling-with-Google-Data-Studio/blob/a31bf2de1ca10db433cf9d0ecb15c3cf4fa882d2/online_sales.zip. This dataset is also used throughout the book to illustrate several examples.

Determine

The target audience for this dashboard is going to be the director of international business. The key objective of this persona is to improve international business by increasing revenue, the customer base, and expanding to more countries. They need to understand how sales are generated from various countries, spot trends and anomalies, and ultimately identify opportunities for expansion. The director has requested a summary dashboard to monitor the year-to-date and month-to-date sales metrics for international countries. Some of the questions that the dashboard should help answer include the following:

What’s the proportion of international business compared to the whole?How have international sales trended monthly and weekly this year?How many sales are generated from new customers versus repeat customers?How many sales do top products generate over time?What’s the number of total sales generated and the number of active customers in the top countries?

In order to answer these questions, we need sales data generated from online stores at a weekly or daily granularity. We also need the country, product, and customer information in order to be able to slice and dice these attributes.

Design

Once we have determined the target audience and key business question(s) that the dashboard needs to cater to, the next step is to design the right metrics and visual representation elements. The key metrics that can serve this simple summary dashboard could be total sales amount, number of customers, average sales per customer, number of orders, and average sales per order. We then need to look at the data and determine whether any data manipulations and data preparation is required to accommodate the metrics identified. In this case, no data preparation is required before connecting the data to Looker Studio. In a real life scenario, you will almost always need to make some changes to your data in order to visualize it in the way you want.

Next, we need to identify the right visualizations based on the chosen metrics and the questions that need to be answered:

The proportion of international business compared to the whole: We can use a pie or donut chart to show the proportion of an attribute value compared to the wholeWeekly/monthly sales trends broken down by new versus repeat customer sales: We can use a stacked bar chart to compare sales by new versus repeat customers over time and add a line to represent the total sales so that the trend is easier to perceiveTop countries by sales and number of customers: A bar chart is a great choice to compare multiple dimension values for easy comparisonTop products by sales: Again, a bar chart is a great fit hereSales trends of top products over time: A simple time series is best represented by a line chartRevenue distribution across different countries: A filled map provides a single-glance view of the geographical presence and approximate ranking of different geographical regions.

Now is the time to build the narrative and create a wireframe for the dashboard. It involves designing the layout of the dashboard and placing the visuals and other components of the dashboard appropriately on the page. It could be a simple sketch on paper, whiteboard, or digital notes application. You can also use spreadsheet applications such as Google Sheets to quickly plug in some mock or sample data and build simple charts. Drawing tools within document and presentation slide applications can also be leveraged for this purpose to some extent.

For higher fidelity wireframes, you can opt for commercial tools such as Adobe Illustrator, Figma, Balsamiq and so on. The following figure shows the wireframe that's hand-drawn on a digital notes app:

Figure 1.6 – Handdrawn sketch demonstrating the wireframe of the example sales dashboard

The narrative involves the user starting at the key high-level metrics at the top left and then looking at the year-to-date sales numbers for the top countries. Next, the user can look into the monthly and weekly sales trends, and finally, delve into top product sales and trends.

As the target audience, The Director of International Business, expects to consume most of the information at a glance, there are no explicit interactive filters that need to be designed. Given that, having a country filter to enable users to select a particular country and look at the metrics per country may be helpful. We can leverage the cross-filtering capability of the reporting tool (Looker Studio) and have the filled map chart serve as a country filter. Selecting a country from the map can filter the remaining charts on the dashboard. Also, combining the weekly and monthly trends into a single chart through drill-down functionality results in a cleaner look and better user experience.

Develop

With the key dashboard components identified and the design complete, it’s time to implement the design and develop the dashboard. The first thing to do is connect the data to the report. Then we consider whether any custom fields and metrics need to be set up within the tool and create them accordingly. We implement each of the visuals as per the components identified in the design step and build out the dashboard following the layout. The UK is excluded from all the charts except the pie/donut charts depicting the proportion of international sales. Also, only the past 12 months of data need to be selected for the entire dashboard to provide the year-to-date view.

The aesthetics of the report is important. Take care to choose the minimum number of colors on the dashboard to have a cleaner look. We ensure that the same color is used to represent an attribute or metric across different charts within the dashboard as much as possible. This is important to avoid confusion. Implement other aspects such as axes, data labels, and legends with consistency throughout the dashboard. Enable cross-chart filtering for mainly the filled map so that the rest of the dashboard metrics and visuals can be filtered for a particular country. Disable it for the donut charts, where it doesn’t make sense.

With all the design elements implemented, we test the dashboard for data accuracy – that metrics are showing up correctly and interactivity, cross-filtering and drill-down, and consistency. Further testing can also be done to ensure performance, data freshness, and security. Finally, publish the dashboard and share it with the users. The following figure shows what the dashboard could look like:

Figure 1.7 – Example sales dashboard built for a UK-based online retail company

We will be following the same approach to build effective dashboards and reports in Part 3 of this book. There, we will go into the details of how to build various charts and implement other dashboard elements.

Summary

In this chapter, we have discussed what data storytelling is and how it’s distinguished from data visualization. We have understood the core elements of a data story and how data stories differ for static and dynamic content. We have learned about the 3-D approach to building data stories for dashboards that comprises three major stages: determine, design, and develop. In the determine stage, we determine the target users of the dashboard, the key business questions that the dashboard needs to address, and the data required to answer those questions. In the design stage, we build the narrative, define the right metrics, and choose the right visualizations and interactivity. Finally, in the develop stage, we actually build the visuals and interactions and choose the colors and other visual aspects. We also deliver the dashboard to the end users and ensure data freshness as part of this final stage.

It is really important to have a good understanding of the data storytelling elements and approach to be able to build compelling data products.

In the next chapter, we will be taking this understanding a step further and will learn about foundational principles of data storytelling, best practices of visualization, choosing the right chart types, and other design aspects.

2

Principles of Data Visualization

Data storytelling is both an art and a science. The art part refers to the story structure and narrative elements that bind data and visual components together, whereas the science part of data storytelling pertains to the foundational principles of design and visual perception and their application. This book largely concerns itself with building data stories through dashboards and reports. It primarily deals with the science aspect of data storytelling. These forms of data presentation provide limited narrative flexibility owing to the dynamic nature of the data they represent. As the data changes over time, the insights it conveys and the story it tells will change accordingly. This makes it difficult to incorporate a rigid narrative. Hence, much of the emphasis is put on design elements such as the chart types, colors, and layout, that constitute the building blocks of storytelling through data, rather than on the narrative elements.

This chapter introduces several guiding principles and key design aspects to consider while building data visualizations. We will cover foundational concepts such as the simplicity of design, principles of visual perception, organization of content, and information accuracy. These principles form the bedrock of any decent visual design. Color plays a very important role in visual design and greatly affects how visual elements are perceived. We will delve into various aspects of choosing the right color schemes based on the specific use cases and audience needs.

For a deeper and more comprehensive study of these ideas, you can peruse other resources mentioned in the Further reading section at the end of the chapter. Additional data visualization concepts, primarily various chart types and gotchas, are discussed in the next chapter, Chapter 3, Visualizing Data Effectively.

In this chapter, we are going to cover the following main topics:

Understanding foundational design principlesReviewing Gestalt principles of visual perception Using color wisely

Understanding foundational design principles

Well-crafted and effective dashboards are built on the foundations of design and visual perception that have been explored and studied for centuries. In contemporary times, Edward Tufte and Stephen Few are recognized as pioneers and the most notable leading experts of modern-day data visualization. Through their prolific work, they have elucidated the many intricacies of visualizing data and information effectively. In this section, we are going to review the following foundational principles and guidelines that form the basis of any good dashboard:

Simplicity of designOrganizing the layoutAccuracy of information presented

Simplicity of design

The single most important guiding principle to building any visual representation of data is simplicity. Achieving simplicity mainly involves removing all distractions away from the data to be represented. In Edward Tufte’s words, it’s avoiding chart junk. Chart junk or non-data ink is anything that doesn’t represent data or information. It can take the form of redundant data, too many colors, three-dimensional (3-D) effects, dark gridlines, and much more. Chart junk interferes with the effective consumption of information. The rest of this section presents some common examples of chart junk and how they can be minimized effectively.

The following screenshot shows two variations of a bar chart depicting the average unit price for each product category. The gradient color and the textured color of the bars in the two charts respectively are distracting and do not add any information to the chart:

Figure 2.1 – Bar charts with distracting uses of color

Use solid colors to represent data and avoid using extraneous effects in an attempt to make a chart more visually appealing. The same applies to the background textures and patterns, as shown in the following screenshot:

Figure 2.2 – Charts with distracting backgrounds

Background images for charts prove to be even more damaging to the eye, making it much more difficult to read and understand the data, as is the case here:

Figure 2.3 – Chart with ineffective use of images

While relevant and beautiful images usually grab the attention of users and it seems like a good idea to include them, a chart or a dashboard is not the right place for them. They distract users away from the objectives of the chart and jeopardize its utility. Use neutral solid colors as backgrounds, or none at all. This particular example showcases an especially bad case of using images, where the image doesn’t add any informational value. While the chart can perhaps be improved in other ways such as better color choice, labels, and so on, images as chart backgrounds seldom add value.

That said, images can aid data storytelling when justly used. Logos are often included in dashboards for branding purposes, while meaningful and relevant icons add value by enabling easier interpretation and increasing visual appeal. Pictographs are also a good example of displaying data as images. They are effective only for a small amount of data, though. The following screenshot shows male and female image icons (sourced from free-vectors.net under the Attribution 4.0 International (CC BY 4.0) license https://creativecommons.org/licenses/by/4.0/) that add context to the donut charts and a pictograph depicting the performance rating with star images (image by rawpixel.com):

Figure 2.4 – Meaningful and appealing use of images

What about gridlines? Are they really chart junk? Judiciously placed gridlines help users to interpret charts easily. If there are too many lines or bars in a chart, adding gridlines may clutter the chart further. Dark-colored and thick gridlines are definitely noisy and considered non-data ink. Adding a large number of gridlines also contributes to chart junk. Use lighter and thin lines sparingly so as to enable the user to read the chart without difficulty while taking care not to clutter the chart. The following screenshot shows a line chart with distracting gridlines on the left and a chart with de-emphasized and useful gridlines on the right:

Figure 2.5 – Use gridlines in charts sparingly and in a non-interfering way

Simplicity also involves showing or encoding only enough information and nothing more. It is often tempting for a dashboard developer to incorporate all the different analyses and details that users can ever find useful into the dashboard and make it all-encompassing. Such a tendency leads to inefficiency and hinders users from achieving their objectives from the dashboard. If the users need to sift through the many visuals and details provided in the dashboard to get their basic questions answered, the dashboard is not simple enough. It is a case of Too Much Information (TMI). A visual—and thereby a dashboard—should contain the right amount of detail and precision to serve the objectives of the target users. The dashboard shown in the following screenshot attempts to provide information about a lot of aspects of Google Ads—campaigns, ad types, search keywords, devices, ad channels, asset types, and more. The level of detail varies, and it has a cluttered look:

Figure 2.6 – Dashboard displaying too much information

Providing appropriate labels and context is key in making a dashboard easily readable. This includes elements such as titles, annotations, legends, units of measurement, and more. Consistency in the font style, font size, legend position, units of measurement, and numerical precision is a key enabler of simplicity in design. Having consistent elements across the dashboard puts a lower cognitive load on users as they do not have to process each element differently.

Removing, hiding, and de-emphasizing the details of less-important data is as essential as highlighting more important data elements. A dashboard where everything is highlighted is one in which nothing stands out. The way various elements in a chart or a dashboard are organized also affects how effective and meaningful the chart can be. The design should serve how we humans naturally perceive information visually.

Organizing the layout

The human eye typically follows the path of Z