🪓Data Journalism Unit 7 – Data Visualization Principles and Best Practices

What's the Big Deal?

• Data visualization transforms complex data into easily digestible visual representations (charts, graphs, maps)
• Enables readers to quickly grasp trends, patterns, and outliers in large datasets
• Enhances storytelling by making data more engaging and memorable
• Facilitates data-driven decision making by highlighting key insights
• Bridges the gap between raw data and human understanding
  • Especially important in an era of big data and information overload
• Empowers journalists to uncover hidden stories and hold those in power accountable
• Helps audiences understand the significance and impact of data-driven stories

Key Concepts to Know

• Data types: categorical, ordinal, and quantitative
  • Categorical: discrete categories (gender, race, product type)
  • Ordinal: categories with a natural order (low, medium, high)
  • Quantitative: numerical values (age, income, temperature)
• Encoding: mapping data to visual properties (position, size, color, shape)
• Scales: mapping data values to visual dimensions
  • Linear, logarithmic, and categorical scales
• Gestalt principles: how humans perceive visual elements as unified wholes
  • Proximity, similarity, continuity, closure, and figure-ground
• Preattentive attributes: visual properties processed quickly by the human brain (color, size, orientation)
• Interaction techniques: enabling users to explore data (zooming, filtering, highlighting)
• Responsive design: ensuring visualizations adapt to different screen sizes and devices

Tools of the Trade

• Spreadsheet software (Microsoft Excel, Google Sheets)
  • Organizing, cleaning, and analyzing data
  • Creating basic charts and graphs
• Programming languages (Python, R)
  • Powerful libraries for data manipulation and visualization (Matplotlib, ggplot2)
  • Customizable and reproducible visualizations
• Business intelligence platforms (Tableau, Power BI)
  • Drag-and-drop interfaces for creating interactive dashboards
  • Connecting to various data sources and real-time updates
• Web-based tools (D3.js, Plotly)
  • Creating interactive and animated visualizations for the web
  • Leveraging web standards (HTML, CSS, JavaScript)
• Graphic design software (Adobe Illustrator, Sketch)
  • Refining and polishing visualizations for publication
  • Creating custom visual elements and layouts

Design Principles that Pop

• Simplicity: focusing on essential information and minimizing clutter
  • Removing unnecessary elements (gridlines, borders, labels)
  • Using clear and concise labels and annotations
• Visual hierarchy: guiding the reader's attention through strategic use of visual elements
  • Emphasizing key data points with size, color, or position
  • Grouping related elements and creating a logical flow
• Color theory: using color effectively to convey meaning and enhance understanding
  • Choosing a color palette that aligns with the data and message
  • Considering color blindness and ensuring accessibility
• Typography: selecting fonts that are legible and appropriate for the context
  • Using a limited number of font families and sizes
  • Ensuring proper contrast between text and background
• Consistency: maintaining a cohesive visual style throughout the visualization
  • Using a consistent color palette, font, and layout
  • Aligning elements and maintaining proper spacing
• Data-ink ratio: maximizing the amount of ink used to display data and minimizing non-data ink
  • Removing unnecessary backgrounds, borders, and decorations
  • Using a minimalist approach to highlight the data

Common Pitfalls to Avoid

• Overcomplicating the visualization with too much information or visual elements
  • Overwhelming the reader and obscuring key insights
• Using inappropriate chart types for the data and message
  • Distorting the data or creating misleading impressions (pie charts for non-proportional data)
• Misusing color, leading to confusion or misinterpretation
  • Using too many colors or colors with conflicting meanings
  • Failing to consider color blindness or cultural associations
• Ignoring the importance of context and annotations
  • Leaving the reader to interpret the data without guidance
  • Missing opportunities to highlight key findings or provide explanations
• Neglecting the target audience and their level of data literacy
  • Creating visualizations that are too complex or technical for the intended readers
• Sacrificing accuracy for aesthetics
  • Manipulating scales or truncating axes to exaggerate differences
  • Cherry-picking data to support a predetermined narrative

Hands-On Techniques

• Sketching: starting with pen and paper to explore ideas and layouts
  • Rapidly iterating on design concepts before moving to digital tools
• Data preparation: cleaning, transforming, and aggregating data for visualization
  • Handling missing values, outliers, and inconsistencies
  • Reshaping data into a format suitable for visualization (long vs. wide format)
• Exploratory data analysis (EDA): using visualizations to uncover patterns and insights
  • Creating multiple chart types to examine the data from different angles
  • Identifying correlations, trends, and outliers
• Prototyping: creating quick and rough visualizations to test ideas and gather feedback
  • Using tools like Excel or Tableau to rapidly prototype designs
  • Iterating based on feedback from colleagues or target audience
• Refinement: polishing the visualization for final publication
  • Adjusting colors, fonts, and layout for optimal readability and aesthetics
  • Adding annotations, labels, and legends to provide context and guidance
• Interactivity: incorporating interactive elements to enable data exploration
  • Adding tooltips, filters, or hover effects to reveal additional details
  • Allowing users to zoom, pan, or select data points of interest

Real-World Examples

• The New York Times' "The Rich Really Do Pay Lower Taxes Than You"
  • Using a line chart to compare tax rates across income groups over time
  • Highlighting the stark contrast between the ultra-wealthy and average taxpayers
• Reuters' "The Rohingya Exodus"
  • Combining maps, charts, and photographs to tell the story of the Rohingya refugee crisis
  • Visualizing the scale and impact of the mass displacement
• The Pudding's "The Largest Vocabulary in Hip Hop"
  • Interactive visualization allowing users to explore the vocabulary of famous rappers
  • Demonstrating the power of interactivity to engage audiences with data
• The Washington Post's "2°C: Beyond the Limit"
  • Using maps and data visualizations to show the uneven impact of climate change
  • Contextualizing data with storytelling and human experiences
• ProPublica's "Miseducation"
  • Interactive database enabling users to explore racial disparities in U.S. schools
  • Empowering readers to investigate educational inequities in their own communities

Ethical Considerations

• Accuracy: ensuring data is properly collected, analyzed, and represented
  • Verifying data sources and methods
  • Providing transparency about data limitations and uncertainties
• Integrity: maintaining objectivity and avoiding bias in data selection and presentation
  • Resisting the temptation to cherry-pick data to support a predetermined narrative
  • Presenting data in a fair and balanced manner
• Privacy: protecting individual privacy and confidentiality
  • Anonymizing sensitive data and using aggregation to prevent identification
  • Obtaining informed consent when collecting personal data
• Accessibility: ensuring visualizations are accessible to all users, including those with disabilities
  • Following web accessibility guidelines (WCAG)
  • Providing alternative text for images and using color-blind friendly palettes
• Transparency: being open about data sources, methods, and limitations
  • Providing access to raw data and methodology
  • Acknowledging uncertainties and potential biases
• Impact: considering the potential consequences of data visualizations on individuals and society
  • Being aware of how visualizations may influence public opinion or policy
  • Taking responsibility for the ethical implications of data-driven storytelling