insights from data visualization - stephen lett (procter & gamble)
DESCRIPTION
This month, we will dive into the world of data analysis and visualization. As data continues to proliferate our lives and work, the question of how to make sense of it and turn it into information and knowledge becomes more and more challenging. At the same time, powerful tools are becoming available to help analysts sift through data and present it in a way that draws attention to key bits of knowledge than can be derived. As such, the skills related to using these tools effectively have become highly sought-after as organizations seek to dig out the treasures hidden in their data troves. Presentation by Stephen Lett (Procter & Gamble)TRANSCRIPT
Insights from Data VisualizationSTEPHEN LETT
19 NOVEMBER 2014
Introduction
Procter and GambleGlobal Business Services, Global Service Manager
15 Years Career 3 Years – Planning and
Manufacturing Business Intelligence Service Manager.
Data Visualization Discussion Points
What is it?Brief overview of what data visualization is and why it is important.
What to Know First?Questions to consider when deciding how to best leverage Data Visualization.
Visualization DesignFew golden tips to remember to make your visualization stand out.
Quick Survey
When was the first line chart and bar graph produced?
1639 1786 1801 1920
William Playfair,Scottish Engineer
1759-1823
Data Visualization Transformation
Reporting
Visualizing
Business Intelligence
Data Visualization WHAT IS IT?
What is Data Visualization?
Simple Definition: Visual Representation of Information
A means of communicating information clearly and effectively through graphical displays
Edward Tuft, 1983
At their best, graphics are instruments for reasoning about quantitative information.
Often the most effective way to describe, explore, and summarize a set of numbers - even a very large set - is to look at pictures of those numbers.
Well-designed data graphics are usually the simplest and at the same time the most powerful.
Why is Data Visualization Important in Business?
Why is Data Visualization Important in Business?
VUCA World Data is the consistent tool that simplifies the VUCA into where we need to
focus.
Translates into COST SAVINGS Getting to the right answers faster is the only differentiator in competitive
markets.
Fosters PRODUCTIVITY and COLLABORATION Good graphical representations of data communicates complex ideas with
clarity, precision, and efficiency.
Getting teams on the same page faster, and individual contributors on a head start for where to focus.
Data Visualization
WHAT TO KNOW FIRST?
3 Factors to Consider:
1.Who’s the “Who”?
2.What’s the “What!”?
3.How to Motivate towards ACTION?
Who’s the Who?
Know Their Background: Experts in the Field/Subject?
Analysts – using the data for projections / transformations?
Management – strategic directions?
Define Consumerization of Data
Instructor Driven
Self-Discovery
Set the Right Expectations
Basic Overview
Category Glance
Deep Discovery in 1 Vector
Communication vs Analysis
Visualization for Communication
Visualization for Analysis
Audience General PublicSenior Leaders
Analysts
Intent Summary & ConclusionsExplain the Situation
Explorations and Observations
Data Consumption Immediate Understanding Required
Complexity SIMPLE Situational Based
Time to Generate Fairly Quick Real-Time**
What’s the Message?
Background / General Knowledge?Summary to Support a Decision?Deep Analysis to Drive a New Activity?Communicating new insights?
Plan Accordingly!
Good Rule of Thumb:
Work on Audience and Message understanding and intent BEFORE building any visualizations!
Motivating Towards ACTION
Based on the AUDIENCE and the MESSAGE Pick the visualization that will lean towards ACTION!- If you can communicate the message clearly and efficiently in a simple sentence, DO IT;- If Data Tables are required, use them – but don’t lean on visual perception alone. Manage the message!
Data Visualization
VISUALIZATION DESIGN
TIPS
Key Steps in Designing your Visualization
Data PreparationClean Data is a MUST:- Eliminate “Noise” (e.g.
nulls, missing values)- Clarify data (full set,
representative sample, etc)
Normalize and Transform Upfront- Aggregate- Filter- Primary / Secondary Keys
Choose Your Graph• Amount of Data
• Type of Data
• Data Relationships
• Conclusion for Audience
Good Design Principles
1. Avoid “Chartjunk”
2. Use Colors Wisely
3. No Misleading Scales
4. Dual Axis Charts are for Experts
Good Design Principle 1 – Chartjunk
What is It? Visual Content that:
Adds little / no value
Serves little / no purpose
Distracts from real data
Examples Shadows / Color Effects
Good Design Principle #2 - Colors
If the point is: Consistent performance at 40%+, which chart uses color most effectively ?
• Use the same color, except when color differences make a difference
• Use a single, neutral background color (if needed at all)
Good Design Principle #3 - Scaling
How are we doing with our budget forecast vs actuals over last 6 months?
Notice the “Y-Axis”? Always NOTE if an axis doesn’t start with 0
Good Design Principle #4 – Dual Axes
Summary
Data Visualization is:
Communicating clearly and effectively through graphics.
Know Your Plan1. Who’s the Who?
2. What’s the “What!”?
3. Motivate to ACT!
Golden Design Tips1. Avoid Chartjunk
2. Use Colors Wisely
3. Beware of Scaling
4. Dual Axes Needed?
Foundational Principles
If the message is simple, keep it simple.
If the message is complex, make it look simple.
Always tell the truth – don’t use graphs to distort the data.
Questions?
Backup Slides
William Playfair Graphs 28
William Playfair,Scottish Engineer
1759-1823
Does it provide insight or understanding that was not obtainable with the original representation (text, table, etc)?
Q3. Is visualization the best way to share the data, show the findings, and/or reveal the insight?
29
Tables Graphs
Data are arranged in columns and rows Data are displayed in relation to one or more axes along which run scales that assign meaning to the values
work best when the display will be used to look up individual values or the quantitative values must be precise.
work best when the message resides in the shape of the data (that is, in patterns, trends, and outliers).