visual analytics best practices

4
White Paper Visual Analytics Best Practices Organizations of all sizes from all industries recognize the importance of giving their employees the ability to conduct everyday data analysis, regardless of technical background. Companies don’t need more staff; they need more access to the information they’ve already got, across all tiers. The better employees understand the data in front of them, the stronger their decisions will be. As a result, tools that efficiently and intuitively analyze all types of data are in demand. Organizations and individual users are gravitating toward visual analytics tools as one of the most effective ways for the everyday user to manipulate data and maximize its insights. WHAT IS VISUAL ANALYTICS? Visual analytics pairs graphical illustrations with specialized interaction to reveal underlying patterns in data. Essentially, it turns dense, confusing data—big or small—into accessible, dynamic, and predictive information. Rather than providing a “rear-view mirror” perspective, visual analytics allow the user to interact with data in real time. It provides insight into circumstances and opportunities as they develop, furthering intelligent decision-making and organizational improvement. The great news is visual analytics isn’t something that’s only available to IT professionals. Today’s tools are designed with the everyday user in mind, giving them access to information traditionally reserved for data scientists and analysts. Imagine that your supervisor asks you to break down how this quarter’s marketing campaigns are performing. Or why the southwest regional sales team is lagging behind the others. Or what percentage of Facebook fans interact with your organization’s posts, and which types of posts garner the largest responses. With a visual analytics tool, you can dig in and analyze these questions in real time. For example, using your company’s Facebook metrics, a visual analyics tool will enable you to create a visualization that plots each post’s level of engagement, as it develops. With a simple click of the mouse, the tool also allows you to drill down on specific types of posts so that you can compare success rates. Within a few minutes, you’ll be able to report back to your supervisor that Facebook fans interact most with posts that include photos, and you’ll be able to support that assertion with numbers and graphics. WHY EXCEL SHEETS AND POWERPOINT DECKS FALL SHORT All organizations are driven by decisions—and the best decisions are driven by data. But data analysis can be confusing or overwhelming for many people; particularly those without a background in statistics. That’s why it’s essential to understand the three main challenges visual analytics users face. Visual analytics has the ability to take data analysis to the next level, but it isn’t a magic bullet. What you know is of very limited value if you’re unable to use clean data, leverage the most powerful visuals, and continually share and refresh your findings. 1) Understanding the data and its structure. On a basic level, you need to be aware of the range of data that exists, what it means, and how different datasets relate to each other. (A dataset is a collection of related data, usually displayed and organized in tabular or graphical form.) If you don’t understand the information with which you’re working, or if there are gaps in your knowledge, you won’t be able to move forward effectively. 2) Presenting the data effectively. After you have digested the available data, you must present it in way that’s manageable, comprehensible, and able to drive action. Otherwise, it’s essentially useless—an inert pile of information. With large, dense, and constantly growing amounts of data, overcoming this particular challenge is becoming simultaneously more difficult and more important. 3) Sharing the data. After you have analyzed and organized the data, you’ll need a way to share it with others. The challenge lies in getting all parties on the same page and in a data analysis mindset. You’ll also need to provide a way for

Upload: logi-analytics

Post on 06-May-2015

251 views

Category:

Data & Analytics


3 download

DESCRIPTION

In this white paper, learn the fundamentals of visual analytics – how to empower non-technical employees to conduct everyday data analysis, why tools like Excel and PowerPoint don't work and how to visualize key insights to convince stakeholders to act.

TRANSCRIPT

Page 1: Visual Analytics Best Practices

White Paper

Visual Analytics Best Practices

Organizations of all sizes from all industries recognize the importance of giving their employees the ability to conduct everyday data analysis, regardless of technical background. Companies don’t need more staff; they need more access to the information they’ve already got, across all tiers. The better employees understand the data in front of them, the stronger their decisions will be. As a result, tools that efficiently and intuitively analyze all types of data are in demand. Organizations and individual users are gravitating toward visual analytics tools as one of the most effective ways for the everyday user to manipulate data and maximize its insights. WHAT IS VISUAL ANALYTICS? Visual analytics pairs graphical illustrations with specialized interaction to reveal underlying patterns in data. Essentially, it turns dense, confusing data—big or small—into accessible, dynamic, and predictive information. Rather than providing a “rear-view mirror” perspective, visual analytics allow the user to interact with data in real time. It provides insight into circumstances and opportunities as they develop, furthering intelligent decision-making and organizational improvement. The great news is visual analytics isn’t something that’s only available to IT professionals. Today’s tools are designed with the everyday user in mind, giving them access to information traditionally reserved for data scientists and analysts. Imagine that your supervisor asks you to break down how this quarter’s marketing campaigns are performing. Or why the southwest regional sales team is lagging behind the others. Or what percentage of Facebook fans interact with your organization’s posts, and which types of posts garner the largest responses. With a visual analytics tool, you can dig in and analyze these questions in real time. For example, using your company’s Facebook metrics, a visual analyics tool will enable you to create a visualization that plots each post’s level of engagement, as it develops. With a simple click of the mouse, the tool also allows you to drill down on specific types of posts so that you can

compare success rates. Within a few minutes, you’ll be able to report back to your supervisor that Facebook fans interact most with posts that include photos, and you’ll be able to support that assertion with numbers and graphics. WHY EXCEL SHEETS AND POWERPOINT DECKS FALL SHORT All organizations are driven by decisions—and the best decisions are driven by data. But data analysis can be confusing or overwhelming for many people; particularly those without a background in statistics. That’s why it’s essential to understand the three main challenges visual analytics users face. Visual analytics has the ability to take data analysis to the next level, but it isn’t a magic bullet. What you know is of very limited value if you’re unable to use clean data, leverage the most powerful visuals, and continually share and refresh your findings.

1) Understanding the data and its structure. On a basic level, you need to be aware of the range of data that exists, what it means, and how different datasets relate to each other. (A dataset is a collection of related data, usually displayed and organized in tabular or graphical form.) If you don’t understand the information with which you’re working, or if there are gaps in your knowledge, you won’t be able to move forward effectively.

2) Presenting the data effectively. After you have

digested the available data, you must present it in way that’s manageable, comprehensible, and able to drive action. Otherwise, it’s essentially useless—an inert pile of information. With large, dense, and constantly growing amounts of data, overcoming this particular challenge is becoming simultaneously more difficult and more important.

3) Sharing the data. After you have analyzed and

organized the data, you’ll need a way to share it with others. The challenge lies in getting all parties on the same page and in a data analysis mindset. You’ll also need to provide a way for

Page 2: Visual Analytics Best Practices

White Paper

everyone to collaborate once they’ve analyzed the data for themselves.

These challenges make traditional methods of data sharing inefficient, frustrating, and sometimes, an utter waste of time. We’ve all pulled our hair out while trying to pull PivotTables out of Excel for weekly status updates, or attempted to come to a group consensus via a lengthy, multi-pronged email chain that leaves participants more confused than they were at the outset. Visual analytics tools make it easier than ever before for non-technical business users to navigate and overcome these challenges. But it’s not enough to have the technology—you need to use these tools effectively. Now, let’s delve into visual analytics best practices. BEST PRACTICE: IDENTIFY YOUR GOAL FOR ANALYSIS What are you trying to accomplish as you analyze the dataset? Keeping this purpose in mind guides the development of each visualization and helps tell a complete story. For instance:

• Will you be comparing over time or across categories?

• Will you be relating variables to one another? • Will you be looking for correlations between two

different datasets? • Will you be comparing a part to a whole (e.g.,

comparing your regional division to the entire sales department)?

• Will you be determining the range across which a dataset falls?

You should also give some thought to the data sources you’ll need to tap into. To address the question at hand, will you need to combine datasets for effective analysis? Often, sources will need to be blended in order to give you the desired 360-degree view. These considerations may seem simple, but moving forward without them jeopardizes your reporting. To effectively incorporate the other best practices that are presented in this white paper, you’ll need to keep the purpose of your analysis firmly in mind at all times so that the many options and capabilities offered by visual analytics tools will work for you, not against you!

BEST PRACTICE: GET MORE MILEAGE WITH DASHBOARDS Visual analytics tools allow you to use dashboards to show multiple datasets at the same time. If you’ve worked with dashboards before, you know that they are especially useful when you are considering and/or comparing related datasets. To maximize dashboards, you can:

• Group several types of visualizations on the same screen to show different aspects of the dataset you’re considering. For instance, a bar graph might quantify each team’s current-quarter sales, a line graph might track each team’s sales over the past two years, and a scatterplot might show number of sales relative to number of ads running. Be sure to place reasonable limits on the amount of data included. Otherwise, your dashboard grows distracting and the power of your insights is lost.

• Tier visualizations by size and position to highlight

the most important analysis. For instance, you can draw attention to the most important and relevant visualizations by making them larger than visualizations that are more tangential to your purpose. If you’d like to know more about maximizing your dashboards, this white paper (Dashboard Best Practices: How to Derive the Most Value from Your Data and Improve Your Business) delves deeper into that topic.

Example of a dashboard showing multiple datasets, with important visualizations distinguished by size. For the purposes of this particular white paper, it’s important to understand that the sharing capabilities of traditional data presentation models differ somewhat from the sharing capabilities of the data presentation models found in visual analytics. Specifically:

Page 3: Visual Analytics Best Practices

White Paper

• In traditional data presentation, dashboards display pre-determined datasets, which are often curated by developers and data analysts. They’re usually used to present and share core metrics; for instance, sales over time by region and by representative. While the everyday user might be able to manipulate the data and analysis displayed, you can’t go beyond what’s presented by bringing in new variables and updates.

• In visual analytics data presentation, you can still

analyze and explore the data and analysis presented. However, these dashboards allow you to move beyond “analysis mode” and into “exploration mode.” Starting from any data point, you can explore new questions, begin fresh analysis, and build different datasets based on what’s most relevant and useful to you. For instance, let’s say that you’ve just noticed that sales are down in North America. A visual analytics dashboard will allow you to dig into real-time data, revealing that the slump can be attributed to a lack of leads and the recent resignation of two sales reps.

When you’re using a visual analytics dashboard, keep these key differentiators in mind so that you can leverage them. Not taking advantage of the ability to explore and analyze developing data in real time is like relying on a printed reference book instead of exploring an interactive, multimedia online database. BEST PRACTICE: USE DESIGN EFFECTIVELY Visual analytics tools enable you to display and share data using graphic representations that you create. However, not all charts are created equal. A graph might look attractive, but that doesn’t mean it presents information in a way that can be easily understood and analyzed. This is where many users go wrong: not using the right chart for the right dataset. Your goal is to use design elements in a way that simplifies analysis and helps you drill into the data. Specifically, think about the type of chart you’re using and consider what it encourages you to focus on. For example, a pie chart is best used to show relative proportions between pieces of data, a line chart is best used to track an item over time, and geographic mapping is becoming essential for spatially-relevant data.

Another design element to consider is color. In particular, keep in mind that similar colors may be difficult to easily distinguish from one another, and that color-blind users may not be able to distinguish between different colored elements; most often, red and green. (If you’re unsure of how to proceed, websites like vischeck.com allow you to test your image files to determine how they might appear to color-blind individuals.) What if you aren’t sure which type of chart or which combination of colors will most effectively present your data? Using visual analytics tools allows you to be in the driver’s seat of data exploration. You have the flexibility to create and consider multiple charts before deciding which one makes the most sense to you and your colleagues. When utilized properly, this capability helps preserve your original analysis goals. BEST PRACTICE: REMEMBER THAT MORE ISN’T ALWAYS BETTER Sometimes, more is distracting. In general, the more focused your analysis is, the greater its impact will be. Try to avoid inundating the user with too much data. For example, let’s say that you’re trying to give your supervisors the information they need to make a pressing decision. You know that bombarding them with piles of information isn’t wise. Forcing them to sift through a long string of emails containing dense explanation will only slow things down, or stop progress altogether. Since a relatively small slice of the dataset you’ve compiled will enable your supervisors to make a decision, focus on including only the elements that will drive action, and on organizing your insights so that they’re clear and understood. Note that we’re not unilaterally downplaying background information. Sometimes, larger and deeper datasets are relevant. This is where knowing your purpose for analysis comes into play: being able to identify when additional information will be distracting or helpful. The concept of “more isn’t always better” also applies to design. Recently, there’s been a tendency to move to simpler designs featuring the use of more white space, more careful use of color, and a flat, clean design. Color and shape shouldn’t distract, but help you absorb and digest information.

Page 4: Visual Analytics Best Practices

White Paper

Example of a simple, clean design that doesn’t distract from the data being presented.

BEST PRACTICE: SET APPROPRIATE LIMITSOne of the most notable advantages of visual analytics is that it enables you to share big data with other users. Sometimes, though, you’ll still want to put limits on the data and analysis certain users can access and/or what they’re able to do with it.

As you’re building and sharing visualizations, think about how you might need to filter that analysis in order to conditionally inform what other users see. If your team needs to keep certain information private, even from other groups within your organization, proactively set limits regarding who sees which projects and which sources are accessible: Corporate data files? Specific files developed by your team? The cloud? You might even want to set different levels of permission for different users within your own team.

Think through these implications before making visualizations available to wide groups of users. Whenlarge amounts of data are in play, it can be all too easy for information you’d rather keep private to slip into outside hands.

BEST PRACTICE: DON’T LIMIT YOUR DATA COMING IN OR GOING OUTThis best practice has less to do with how to use a visual analytics tool, and more to do with what to consider when choosing one. That’s because the quality, form, capabilities, and range of a tool are just as important as how you use that tool once it’s in your hands.

Presenting your data in a clean, visually appealing, and easily-understood way is a crucial first step, but even the best visualizations, charts, and dashboards are of limited

use if you aren’t able to deploy and share them effectively. To make sure that your visual analytics data isn’t constrained in terms of reachable audience or functionality, here are four recommendations to keep in mind:

1) Choose a web-based visual analytics tool.Specifically, choose one that is licensed to be distributed to unlimited end-users without additional cost. This will provide you with an easy way to deploy and maintain data, and to share it with as many people as you desire.

2) Choose a visual analytics tool that operates on multiple devices. Don’t restrict yourself to your computer or laptop. Also, look for a tool that will operate within multiple operating systems without requiring additional software.

3) Choose a visual analytics tool that has the ability to export data. If you want to present your data in a different format (for instance, in a PDF, Word, or Excel document) this capability could save you quite a bit of work and time.

4) Choose a visual analytics tool that can be used within other applications. It’s useful if you can take your tool’s content and functionality and embed it within your own application. This is comparable to embedding Google Earth within your company’s application. Whether you’re working alone, with others, or even presenting data to a group, it’s limiting (not to mention frustrating) if you have to keep switching back and forth between your visual analytics tool and another application.