Back to Subject
Practice MCQ

Unit 5 - Notes

INT374

Unit 5: Designing and Enhancing Visual Reports

1. Dashboard Design Framework

The creation of an effective Power BI report relies less on technical proficiency and more on understanding design principles and user psychology. A robust framework ensures the report answers business questions efficiently.

Core Design Principles

  • Clarity: The data must be legible. Avoid "chart junk" (unnecessary gridlines, 3D effects, excessive labels).
  • Consistency: Use the same font families, color palettes, and terminology across all report pages.
  • Context: Data requires context to be meaningful. A number (e.g., "$5M Revenue") is useless without a comparison (e.g., "vs. Last Year" or "vs. Target").

The MAD Framework

A common strategy for structuring analytic applications:

  1. M - Monitor: High-level KPIs and gauges. This answers "What is happening right now?" Usually placed at the top or on the first page.
  2. A - Analyze: Interactive charts (bar, line, scatter) that allow users to slice and dice data. This answers "Why is it happening?"
  3. D - Detail: Granular tables and matrices. This answers "Who/What specifically is involved?" Usually placed at the bottom or on drill-through pages.

Visual Hierarchy and Layout Patterns

  • The Z-Pattern: Users typically scan a page starting top-left, moving horizontally, then diagonally down to bottom-left, and finally across to bottom-right.
    • Top-Left: Most critical KPI (e.g., Total Revenue).
    • Center: Main trend or comparison charts.
    • Bottom: Detailed data tables.
  • The F-Pattern: Common in text-heavy dashboards; users scan the top heavily and read less as they move down the page.

Color Theory in Data Visualization

  • Sequential Palettes: Variations of a single hue (e.g., light blue to dark blue) for continuous data (low to high values).
  • Diverging Palettes: Two contrasting colors with a neutral midpoint (e.g., Red-White-Green) for positive/negative performance against a target.
  • Categorical Palettes: Distinct colors used to separate categories (e.g., Products, Regions).
  • Accessibility: Avoid Red/Green combinations for critical distinctions due to color blindness; utilize symbols or icons in addition to color.

2. Sketching the Layout

Before opening Power BI Desktop, the layout should be planned (wireframed) to minimize rework.

The Wireframing Process

  1. Define Audience and Scope: Who is reading this? Executives need summaries; analysts need details.
  2. Grid System: Divide the canvas into a grid (e.g., 2x2, 3x3) to align objects neatly.
  3. Low-Fidelity Sketching: Use pen and paper, a whiteboard, or tools like PowerPoint/Balsamiq to draw placeholders.
    • Do not focus on specific data values.
    • Focus on which visual type best represents the metric.
    • Plan space for slicers and navigation buttons.

Layout Considerations

  • White Space: Essential for reducing cognitive load. Do not overcrowd the canvas.
  • Navigation: Determine if the report requires bookmarks, page navigation buttons, or drill-through capabilities.

3. Report Pages & Objects

Power BI Desktop treats the "Report View" as a canvas where various objects interact.

Page Setup

  • Canvas Settings: Accessed via the Format pane (wallpaper icon).
    • Type: 16:9 (Default), 4:3, Letter, or Custom (useful for long-scrolling dashboards or mobile layouts).
    • Background: Can apply images or colors. Transparency must be adjusted to see the background.

Static Objects

These elements add structure and context but do not visualize data directly:

  • Text Boxes: Used for report titles, explanatory notes, or dynamic titles (using "Value" integration).
  • Images: Company logos, icons, or background design elements.
  • Shapes: Rectangles and lines used to group related visuals (containers) or create visual dividers.

The Selection Pane

  • Layering (Z-Order): Accessed via the "View" tab.
  • Allows the developer to determine which objects sit on top of others.
  • Crucial for overlaying text on shapes or creating custom navigation menus.
  • Objects can be hidden/unhidden here (essential for Bookmark logic).

4. Visualizations

Basic Chart Types

  1. Column and Bar Charts:
    • Use: Comparing categorical data.
    • Column: Better for time-series (e.g., Sales by Month) or few categories.
    • Bar: Better for long category names (e.g., Sales by Product Name).
    • Stacked vs. Clustered: Stacked shows part-to-whole; Clustered facilitates direct comparison.
  2. Line and Area Charts:
    • Use: Showing trends over continuous intervals (Time).
    • Area Chart: Emphasizes the volume/magnitude of the trend.
  3. Pie and Donut Charts:
    • Use: Showing composition (Part-to-Whole).
    • Best Practice: Use only when categories are few (<5). Avoid if values are similar, as the human eye struggles to compare angles/area.
  4. Card Visuals:
    • Use: displaying a single, critical number (KPI).

Advanced Visuals

  1. Scatter Plot:
    • Use: Showing the relationship (correlation) between two numerical variables (e.g., Sales vs. Discount %).
    • Play Axis: Adds a time dimension, animating the bubbles over time.
  2. Waterfall Chart:
    • Use: Understanding how an initial value becomes a final value through a series of positive and negative changes (e.g., Gross Revenue Net Income).
  3. Funnel Chart:
    • Use: Visualizing a process with stages (e.g., Sales Pipeline: Lead Qualified Proposal Close).
  4. Gauge and KPI Visuals:
    • Gauge: Circular visualization of progress toward a goal.
    • KPI: Displays value, trend line, and status color (Green/Red) relative to a target.

Trend Lines & Forecasts

These features are found in the Analytics Pane (magnifying glass icon) of a selected visual.

  • Trend Lines: Automatically calculate and visualize the general direction of data (Linear, Exponential, Logarithmic).
  • Constant Lines: Add specific reference lines (e.g., Target Line at $1M).
  • Min/Max/Average Lines: Dynamic lines based on the data displayed.
  • Forecasting:
    • Available primarily on Line Charts.
    • Uses exponential smoothing to predict future values based on historical data.
    • Configuration: Set forecast length (e.g., next 3 months), confidence interval (95% or 99%), and seasonality (points per cycle).

Table & Matrix Visuals

  • Table:
    • A standard 2D grid of data. Best for detailed lookups.
  • Matrix:
    • Equivalent to an Excel Pivot Table.
    • Supports Row Headers and Column Headers.
    • Drill-Down: Users can expand hierarchies (e.g., Year Quarter Month) using the +/- icons.
    • Stepped Layout: Controls whether sub-levels are indented in one column or occupy their own columns.

Map Visuals

  1. Bubble Map (Basic Map): Places circles on a map. Size of the circle represents a measure (e.g., Sales volume). Requires precise geospatial data (Lat/Long preferrable to City Names to avoid ambiguity).
  2. Filled Map (Choropleth): Colors geographic regions (States, Countries) based on data density.
  3. Shape Map: Uses custom JSON files (TopoJSON) to map non-standard regions (e.g., Warehouse floor plans or custom sales territories).

Report Slicers

Slicers are on-canvas visual filters that allow users to narrow the data displayed.

  • Format Options:
    • List: Standard checkboxes.
    • Dropdown: Saves space.
    • Tiles: Horizontal buttons (good for mobile).
    • Between/Before/After: Used for Date fields (includes a slider).
  • Sync Slicers:
    • Located in the "View" tab.
    • Allows a slicer on "Page 1" to filter visuals on "Page 2" automatically.
    • Can also hide the slicer on subsequent pages while maintaining the filter effect.

5. Filtering Options

Power BI employs a distinct hierarchy of filtering logic.

The Filters Pane

Allows developers to configure filters that don't take up canvas space (unlike slicers).

  1. Visual Level Filter: Applies only to the selected chart.
  2. Page Level Filter: Applies to all visuals on the current page.
  3. Report Level Filter: Applies to every page in the file.

Top N Filtering

Used to display only the highest or lowest performing items in a category.

  • Implementation:
    1. Select a visual (e.g., Bar chart of Products).
    2. Open the Filters Pane.
    3. Expand the "Product" category.
    4. Change "Filter Type" from Basic to Top N.
    5. Enter the value (e.g., 10).
    6. Drag the metric (e.g., Total Sales) into the "By value" field.
    7. Click Apply Filter.

Conditional Formatting

Enhances tables, matrices, and charts by dynamically changing colors or icons based on cell values.

Types of Formatting:

  1. Background Color: Cells change color gradient (e.g., White to Blue) based on value magnitude.
  2. Font Color: Changes text color (e.g., Red text for negative numbers).
  3. Data Bars: Adds horizontal bars inside table cells (mini bar charts) to visualize relative size.
  4. Icons: Adds symbols (Traffic lights, Up/Down arrows, Flags).

Setting up Rules:

  • Color Scale: Based on lowest/highest values in the column.
  • Rules: Custom logic (e.g., If value >= 0 and < 50% then Red, If value >= 50% then Green).
  • Field Value: Uses a DAX measure that outputs a Hex Color Code (e.g., "#FF0000") for maximum programmatic control.