Data Storytelling

Specificity is the Soul of Data Narrative

The folks in the front of the room stared with a forced intensity at (what must have been) the 23rd straight slide showing data about website performance. Their glazed eyes would have been entirely evident if the speaker wasn’t so intently focused on pointing out the change in bounce rate between August and July. In the back of the room, Brian wasn’t able to summon the energy to care. The gentle hum of laptops, dim lighting, and endless onslaught of data practically begged his mind to wander...

specificity.jpg

Specificity is the soul of narrative

This is a frequently-repeated lesson from John Hodgman's excellent podcast Judge John Hodgman. His fake Internet courtroom demands that its litigants share specific information and stories to bring their arguments to life.

Unfortunately, this lesson is often lost when people use data to communicate. Which is not to confuse detail for specificity. Detail — at least in the data communication context — simply means the access to more and more granular data. Specificity requires something more: delivering information that is familiar to your audience, letting them connect with the subject matter at a more personal level. The data is no longer an abstraction, it is something tangible and real.

How do we deliver more specificity in our data stories? Here are three ideas:

  1. Remind your audience of the people behind the data

  2. Begin with an individual story

  3. Explore individual patterns and behaviors

1. Remind your audience that we are talking about individual people or things.

Data is an imperfect reflection of activity in the real world. You want to find ways to emphasize the connection between real people and the data points shown on the screen. A few examples:

 Use icons as a subtle reminder that we are talking about people

Use icons as a subtle reminder that we are talking about people

 Use images of people to humanize the data

Use images of people to humanize the data

 Use individual components (people) to compose the visualizations. A tradition bar chart is transformed into a stack of the individual units.

Use individual components (people) to compose the visualizations. A tradition bar chart is transformed into a stack of the individual units.

In one memorable meeting, I was demonstrating our workforce analytics solution to a prospective client. I was showing the distribution visualization (above) and was careful to roll over individual people to help explain its meaning. As I was highlighting an employee with 40 years of experience at their company, an executive burst out: “Wait a second, that woman was my elementary school teacher.” The data came to life for him that day.

2. Begin with individual stories before showing the big picture.

One of the all-time best specificity-is-the-soul-of-narrative visualizations is the Gun Deaths visual created by Periscope. Take a moment to experience it.

 To create emotional impact from the data, the designer starts this visual by showing one gun death at a time.

To create emotional impact from the data, the designer starts this visual by showing one gun death at a time.

 Gradually the animation speeds up until the viewer understands the terrifying weight of the many lives cut short.

Gradually the animation speeds up until the viewer understands the terrifying weight of the many lives cut short.

Your data story may be on a more banal topic, but there are still ways to show the individual stories. What does a prototypical conversion in your sales pipeline look like? What is the financial impact of an individual patient going to an abnormally expensive healthcare provider?

3. Provide your audience with the ability to dive into many individual patterns and behaviors.

One compelling anecdote may hook your reader; the ability to see many stories can provide a powerful tool for analysis.

A long time ago we introduced the concept of customer flashcards — visualizations that tell the story of individual people or things, create a language for reading behavior patterns, and the opportunity to flip through many of these visuals. Finding patterns doesn’t have to be the exclusive domain of machine learning — as humans, we are pretty good at seeing and interpreting patterns ourselves. 

Here’s an example from a project we did to see patterns of online learning. Once we found an effective way to show how students took courses, we quickly identified common behaviors that would have been lost in the typical summarization of data. 

flashcards.png

Data storytelling is still finding its fundamental principles and discovering how effectively impact readers. Bringing specificity into these data stories may just be a bedrock principle that we can adopt from a wise Internet judge.

Education Leaders Embrace Data Storytelling

STAT-conference.png

The Data Storytelling Revolution is coming to the K-12 Education world -- in its own unique way. Two days at the annual National Center for Education Statistics STATS DC Data Conference in Washington DC gave me an up-close view of how education leaders were using data to drive policy and understanding school performance. This insiders view was thanks to an invitation by our partners at the Public Consulting Group, one of the leading education consulting practices in the country.

After attending a handful of presentations and hanging out with industry experts, here are a few of my impressions:

Education leaders have a fresh energy about data visualization and data storytelling.

To start with, the conference was subtitled: “Visualizing the Future of Education through Data”. To back this up, the program featured more than a dozen presentations about how to present data to make an impact. There was good-natured laughing and self-flagellation about poor visualizations, and oooh's and aaah's at good visualizations. There was also a genuine appreciation for how important it is to “bridge the last mile” of data to reach important audiences.

IMG_20180730_112920.jpg

Unsurprisingly, Educators understand the need to reach and teach their data audiences.

For many of the attendees, their most important data audiences (teachers, parents, school administrators) are relative novices when it comes to interpreting data. There was a general appreciation that finding better ways to communicate of their data was paramount. The old ways of delivering long reports and clunky dashboards wasn’t going to suffice. The presenters emphasized “less is more” and the value of well-written explanations. I even ran into a solution vendor committed to building data fluency among teachers.  This sincere sensitivity to the needs of the audience isn’t always so prevalent in other industries.

Data technologies and tools take a backseat to process, people, and politics.

On August 20th and 21st, I’ll see you at the Nashville Analytics Summit. When I do, I bet we’ll be surrounded by vendors and wide-eyed attendees talking about big data, machine learning, and artificial intelligence. Not in the Education world. After the lessons of No Child Left Behind and years of stalled and misguided data initiatives, Education knows that successful use of data starts with:

  1. Getting people to buy-in to the meaning, purpose, and value of the data;
  2. Establishing consistent processes for collecting reliable data;
  3. Navigating the political landmines required to move their projects forward.

The Education industry is more focused on building confidence in data, than in performing high-wire analytical acts.

Education has not yet found the balance between directed data stories and flexible guidance.

I sat in on a presentation by the Education Department where they shared a journalism-style data story that revealed insights about English Learners. There website was the first in a series of public explorations of their treasure-trove of data.

Our_Nation_s_English_Learners.png

On the other extreme, the NCES shared a reporting-building engine for navigating another important data set. On one extreme, a one-off static data story; on the other, a self-service report generation tool. The future is in the middle — purposeful, guided analysis complemented by customization to serve each individual viewer. The Education industry is still finding their way toward this balance.

self-service-bi.jpg

 

Every industry needs to find its own path to better use of data. It was enlightening for me to see how a portion of the K12 Education industry is evolving on this journey.

Data Storytelling: What's Easy and What's Hard

Putting data on a screen is easy. Making it meaningful is so much harder. Gathering a collection of visualizations and calling it a data story is easy (and inaccurate). Making data-driven narrative that influences people...hard.

Here are 25 more lessons we've learned (the hard way) about what's easy and what's hard when it comes to telling data stories:

Easy: Picking a good visualization to answer a data question
Hard: Discovering the core message of your data story that will move your audience to action

Easy: Knowing who is your target audience
Hard: Knowing what motivates your target audience at a personal level by understanding their everyday frustrations and career goals

Easy: Collecting questions your audience wants to answer
Hard: Delivering answers your audience can act on

Easy: Providing flexibility to slice and dice data
Hard: Balancing flexibility with prescriptive guidance to help focus on the most important things

Easy: Labeling visualizations
Hard: Explaining the intent and meaning of visualizations

Easy: Choosing dimensions to show
Hard: Choosing the right metrics to show

Easy: Getting an export of the data you need
Hard: Restructuring data for high-performance analytical queries

Easy: Discovering inconsistencies in your data
Hard: Fixing those inconsistencies

Easy: Designing a data story with a fixed data set
Hard: Designing a data story where the data changes

Easy: Categorical dimensions
Hard: Dates

Easy: Showing data values within expected ranges
Hard: Dealing with null values

Easy: Determining formats for data fields
Hard: Writing a human-readable definition of data fields

Easy: Getting people interested in analytics and visualization
Hard: Getting people to use data regularly in their job

Easy: Picking theme colors
Hard: Using colors judiciously and with meaning

Easy: Setting the context for your story
Hard: Creating intrigue and suspense to move people past the introduction

Easy: Showing selections in a visualization
Hard: Carrying those selections through the duration of the story

Easy: Creating a long, shaggy data story
Hard: Creating a concise, meaningful data story
 
Easy: Adding more data
Hard: Cutting out unnecessary data

Easy: Serving one audience
Hard: Serving multiple audiences to enable new kinds of discussions

Easy: Helping people find insights
Hard: Explaining what to do about those insights

Easy: Explaining data to experts
Hard: Explaining data to novices

Easy: Building a predictive model
Hard: Convincing people they should trust your predictive model

Easy: Visual mock-ups with stubbed-in data
Hard: Visual mock-ups that support real-world data

Easy: Building a visualization tool
Hard: Building a data storytelling tool

What's in a Juicebox: Connected Visuals

The ability of an excel novice (i.e. me) to use a pivot table is basically naught. My ability to manipulate data does not exist, and yet I work for one of the most forward-thinking data presentation companies! Nevermind why I was hired, I quickly learned how to use a Juicebox application because Juicebox is designed with the everyday end user in mind. We have tackled the problem of data delivery to both analytical and non-analytical groups. In this post, I want to chat about one of the features that make that possible: connected slices. What is a slice? A slice is a Juice term for a data visualization within a section of Juicebox application.

Screen Shot 2018-03-14 at 11.57.23 AM.png

I have mentioned before that Narrative Flow is important to Juicebox. Our applications are web-based and users expect to move and navigate from top to bottom, like when interacting with a webpage. Part of that movement from top to bottom in Juicebox means that as the user is making selections within the application, those selections should not only carry down the page but that they should also inform the visuals that follow.

We strive to be the world's best platform for telling data stories and because of that connecting our visuals together is vital. When someone makes a selection in the topmost slice, it places a filter on the data and the selection they make. This filter helps the user narrow down their selection and drill into the data.

Much of the problems with static reports and dashboards is that they only give the user a top-level view of his or her data. Traditional solutions do not provide the ability to drill further to discover what factors could be driving the data. In essence, today's charts, dashboards, reports, and BI solutions give the user a snapshot and not the whole story. 

Curious to see what else is included in Juicebox? Check out some of these posts highlighting other unique features:

Getting unstuck: Give your data a jumpstart

It’s a predicament that we’ve seen many times over: your data is stuck. You’ve tried some reporting through some Excel pivot tables, or you’ve messed around with a Tableau trial, but felt like there wasn’t enough engaging content to get your users excited. Rationalizing why you can’t get your data to be impactful for your business, you think things like, “maybe my users are talking about the data but I just don’t know about it” or “maybe the data isn’t structured in a way that allows for valuable insights to be extracted from it."

If you’re sitting there thinking that your mind is being read by our artificial intelligence, you’re wrong. It's because at Juice we have seen this scenario played out too many times and we’ve made it our mission to make these issues a thing of the past. What you need to do is give your data a jumpstart.

Here’s our suggested plan of action for getting your data unstuck and giving it the jumpstart it needs:

1. Get your data into a readable structure.

  • The first row of your data should always represent the column’s title
  • Columns should contain the same type of values, respectively

  • Each row should represent a case or a single instance within the data and should contain a date of when that data was collected. This means that two different rows in the data can represent the same entity with data collected for it at different points in time.

  • As a consequence of the rule above, the data should include a row identifier column that can be repeated to indicate that different rows of data are representing the same entities.

  • Make yourself a metadata sheet (also commonly known as “data definitions”) that you and other users of the data can refer to.

Here are some simple example data & metadata using the principles above.  

2. Present your data in hierarchical manner catered to specific audiences.

  • Give your audience a call-to-action, let them know why the data is important and why they should care.

  • Begin with presenting high-level key metrics. Think about what the most important numbers are you to your intended audience(s).

  • Give your audience the option to select a few different categories in which to segment and parse-out those important numbers. Doing this will allow your audience to drill-down in the data to get from a high-level to a granular level.

  • Allow your audience to take the data they have drilled down to with them. This could be one row of data out of the thousands they started with at the high-level.  

Here’s an example of this data presentation flow.

3. Engaging your audience in data discussions

  • This one is self-explanatory: talking about the data with others is the best way to squeeze the value from it.

Here’s an example of effective data discussions.

Sounds like a good plan of action, right? If you're still not sure what your next steps should be, we’re here to help.

We’ll work with you to get your data in a structure that makes it valuable, or even create data for you. We’ll build you a data story with that data that helps you and your users understand the data so that you can turn data insights into business actions. We’ll get your users engaged in data discussions and app design feedback so that you know they’re engaged with the data and you know how valuable they perceive the app to be. So drop us a line, we’re here to help.

What's in a Juicebox: Narrative Flow

Do you remember when Facebook first launched its News Feed? I do. I can recall complaining with friends that it was "too stalker-ish" (we were right) and that "no one was going to use it" (we were wrong).

Today it's hard to imagine Facebook without its signature News Feed, but for a long time I did not care for it at all. That's because once people get used to a certain way that things operate on the web, they don't like it when those practices are changed. Not only do they dislike change, but they come to expect certain design practices online. Take for example the horizontal navigation bar found across the tops of web pages: around 88% of websites have its main navigation panel there on every page. Nine times out of 10, that's the first place a user will look when attempting to navigate through a website.

Another piece of design experience that all users expect is vertical narrative flow. Web pages flow from top to bottom and as you want to learn more about a company, product, piece of news, etc. you scroll down.

Juicebox is not unique among BI solution to offer a design layout that flows from top to bottom. Infographics and web pages have taught us that people want to read data the way the same way that they read text. However, Juicebox has a special ability to seamlessly connect the narrative flow, dynamic textual content, and complex filters to give users an effortless experience while navigating data.

interactive-slices.gif

Narrative flow is essential to the experience of Juicebox and our user-centered design. As a user interacts with a Juicebox application, they are continually making decisions about what they would like to see in the data, what relationships are most important to them between segments and tables, and what details they need to make informed decisions. All of these complex interactions are done behind the scenes as Juicebox provides that infographic-type feel in a web-based format.

gif-Narrative Flow.gif

Can you think of a situation where you need to deliver data to an audience of users? Maybe it is customer reporting or a data model that needs a user interface for people who are not data savvy. In recent months we have made it even more easy for you to get started with Juicebox. Through our Guided Design Process you can see your data in Juicebox and give access to 10 users so that they can experience the ease of navigating through your data in a narrative flow. To learn more about Juice's Guided Design Process, check out our resources page.

What's in a Juicebox: Discussions

What good is information if it cannot be shared and discussed? One of the founding pillars of Juicebox is communication; we aim to allow users, regardless of their familiarity with data analysis, the ability to easily identify and discuss important data points.

In taking on this challenge of what we call "The Last Mile of Business Intelligence", the question we must constantly ask ourselves is, "How do we make starting a conversation around data as easy as sending a text message?"

In order to solve for this, we have taken the knowledge gained from our 11+ years of designing and creating custom data applications and created an interactive data storytelling platform that helps everyday information workers make smarter decisions. Our goal for Juicebox is to reinvent the way people discuss and communicate data in the workplace and to their customers.

Our Discussions feature within Juicebox does just that by enabling those conversations around data in a method that is intuitive, quick, and effortless, especially compared to traditional processes. In the past when an insight was discovered within a spreadsheet, an analyst would have to send a report to a decision maker and ask him or her to review the finding. Not only was this process clumsy and time-consuming, the analyst and the decision maker were often on different planes in terms of data skill level. With Discussions, those conversing over the data can take a snapshot of the visual, mark it up, and download it in order to ensure the most relevant and important information is being shared. 

If you're interested in having conversations around your data, we would love to talk with you. Send us a message at info@juiceanalytics.com or click below to tell us more about what you're looking for. 

Data Storytelling Workshops, Part 2: Data Story Showcase

This is part two in a series on sharing Juice’s data storytelling method at various workshops around the United States. In part one, we talked about the highlights of teaching business professionals how to build insightful data stories in under an hour. Here we’ll showcase a data story that was built by one of our own Juicers who attended a workshop.

When creating data stories at Juice’s data storytelling workshops, we always start with a set of data from Nashville’s Open Data project. There are an infinite amount of data stories that can be created from a set of data like this that contains information on construction permits, location, cost, and type of building permit. For our data story prototype, we decided we wanted to know where to find construction projects for multi-family housing so that we could determine where to best build a cool new coffee shop.

giphy (12).gif

When creating our data story, it was important to us that we first give it a title. There are many different strategies that you can utilize when coming up with a title, but we decided to start with a simple yet profound question that would ultimately be answered by the end of the data story.

We find that in order to ensure that the data story being created is coherent and focused, it’s crucial to determine how each visualization contributes to the overall goal of the story. In this example, we wanted to have a high volume of people who enjoy a good cup of coffee and would be likely to visit our coffee shop. To display this in our data story, we would zoom into the map to browse areas around Nashville that are sized by the number of multifamily home projects currently underway in a given zip code.

Once we had selected a zip code that had a business-sustainable number of multifamily projects underway, we also wanted to check to see which way the number of projects in that zip code has been trending over the past 5 years. After seeing that the volume of projects in our zip code of interest had been positively trending over the past few years, we would search through the table at the bottom to find the largest one of these projects to build our coffee shop near so that we can maximize our chances for success.

IMG_8506.JPG

Want to know more about our data storytelling process? Send us a message! We're always happy to share our methodology or to answer any questions.

 

Data Storytelling Workshops, Part 1

The Juice team has been traveling around from conference to conference showcase our method of quickly and easily creating data stories from a data set. We got the opportunity to utilize Nashville’s Open Data project to source the data we used for the workshops. Attendees were divided into several groups and given the option to choose between several personas for whom they would build their data story. By focusing on a particular type of user’s goals, attendees were easily able to create questions that should be answered by the data. These questions or “goals” for their data story were written out on sticky notes by each group member and were shared with the entire group.

Juice Werkshops GIF 1.GIF

Once the goals for their stories were distilled into just 3 questions total, attendees chose metrics and dimensions that would function to best answer the questions that would achieve the goals that their personas wanted from the data story. Making decisions about building effective data stories typically take hours if not days. We were able to accomplish this in less than an hour and saw attendees leave with a full understanding of how a great data story is built for a particular audience or user.

Some workshop moments that were captured can be found in the 30-second video below:

Stay tuned for part two, in which we will showcase a data story that was created by one of our attendees. You won’t want to miss what he created in just under an hour with Juice’s guidance!

If you can't wait and want to see how you can start making your own data stories with Juice, send us a message using the button below.

The Art of Data Storytelling: Structure

This is the second in a series of posts on The Art of Storytelling, a video series from Pixar that shares its storytelling methodology. In this post, we will be examining how the lesson on Story Structure can be applied to data storytelling. For part one on storytelling and character, click here

Introduction to Structure

While traditional storytelling and data storytelling are not identical mediums, there is quite a bit of overlap between the two, and many of the best practices for one can be applied to the other. Take for example the idea of structure when it comes to storytelling. Structure, or in simpler terms, “what do you want the audience to know, and when?” is hugely important when it comes to the practice of data storytelling.

It may seem counterintuitive to consider modeling your data presentations after traditional storytelling structure. After all, storytelling is an inherently subjective act. The storyteller is crafting something that helps the audience learn about a theme that the storyteller finds important, and consequently a moral that should be learned. Applying this to data can seem like enemy territory for analysts who feel that their job in presenting data is to “let the data tell the story.” It’s important to note, however, that the data doesn’t have an opinion on what is important. For example, I was speaking to an HR Analytics team recently and it was clear to me that they wanted to use data to share important lessons with the business. It was less clear that they felt empowered to do so because they felt the data should speak for itself. Data often needs a voice to give it meaning.

When creating the structure of your data stories, keep in mind that it often takes a while to get to the structure that works best for what you are trying to accomplish. That is why it is important to create something ‒ even in a rough form ‒ and get it in front of people who will give you feedback. Does it resonate and connect with the audience ‒ or is it more like the unpopular original structure of Finding Nemo? Without this knowledge, you’re more lost than Dory and Marlin ever were.

Story Beats & Story Spine

An effective way of organizing story structure is by utilizing story beats, the most important moments in your story, and story spine, a pattern into which most stories can fit. While your data story most likely won’t open with “once upon a time…” and end with “and ever since then…” the lesson can still be applied. Using a structure that is broadly familiar to audiences and hitting familiar story beats will help ensure that a data story leverages the hooks that storytelling already has in people. Your audience is looking for certain things in a data story, just like they would in a Pixar film. Who or what are the key players? What’s the conflict? How can it be resolved? Utilizing these when appropriate will make your data stories much more effective.

Act 1

The first act of a film serves to introduce the audience to a protagonist, establish the setting, provide information into how the characters’ world works, and introduce an obstacle that sets the rest of the story in motion.

In traditional dashboards and reports, this information is often missing and leads to users not knowing where to start. If your audience is going to go on a data adventure with you, they should start off by caring about the situation that exists. Data stories should start with a high-level summary that then lets users progressively and logically drill into more complex details and context.

Act 2

Pixar states that the second act of a story as “a series of progressive complications.” My favorite way of describing act two is “the part of the story in which you throw rocks at your characters.” Either way, what happens in the next part of your data story is clear: addressing conflict.

When it comes to data stories, act two is the back-and-forth exploration of the problem. In the traditional story spine they refer to it as “because of that…”; for analytics we call it “slicing-and-dicing.” Throughout act two of your data story you are showing your audience the drivers of problems and identifying any outliers.

Act 3

In traditional storytelling, the third act is the part of the story where the main character learns what she truly needs, as opposed to what she thought she wanted. The character has gone on a transformation along the course of the story, and that is evidenced in the final act.

This is much harder to pull off in data storytelling. In data storytelling, I believe the protagonist is the audience. Much like the main character, the audience needs to be transformed and understand something new and important. A satisfying story is when a problem is fixed and the world is set right in some way. Great data stories deliver that change -- but to do so they need to do more than change the audience’s perspective. They need to make the audience act on, not just discuss, this transformation.

Advice

The best bit of advice from the Pixar storytellers is simple: work backwards. This is how we do it at Juice: we consider what is the endpoint, the change or impact that we want to make on the audience, and then craft the story that can help get us there.

We know that crafting data stories can be a challenging process, and that’s why we’re here to help. If you’d like to talk to us about how we create data stories for organizations like the Virginia Chamber of Commerce, send us a message at info@juiceanalytics.com or click the link below.