Data Visualization

Choosing the right visualization is a balances science and creativity, because you will want to accurately convey data and engage your audience with well design charts. 
It is important to choose visualization styles that are consistent with how people process information, while also creating a visualization that  captures viewer attention and tell a story. The choice between a column chart or a pie chart isn't always clear and depends on whether you want to explain the data or let the viewers explore the data and see patterns and relationships.   

The information under this tab will help you to understand what considerations are necessary when planning and designing a visualization. 

Gestalt Psychology is an approach to understanding how visualization information is perceived and organized by the mind.  In the context of data visualization the gestalt approach anticipates that views of visualizations will first focus on the whole of the visualization, identifying general patterns and how the components are organized and related to each other and gradually analyzing specific components of the visualization like single values.

The Gestalt approach is not without criticism but remains a useful way of understanding how people interact with data visualizations, and so can be used to better inform decision making when choosing a visualization type.

The principals of Gestalt Psychology are:

• Similarity : Figures that are similar in shape, color, size, or form are perceived as belonging to the same category.

Using colors, shapes, or patterns consistently in charts will help viewers understand that certain data points belong to the same group. Using different colors or symbols can help distinguish between categories or datasets.

For example, in a line chart with multiple data series, using different colors for each line will make it easy for users to associate similar-colored lines with the same category.

•  Proximity : Elements that are close to each other tend to be perceived as related or grouped together.

In charts or graphs, placing data points or objects closer together helps viewers understand relationships between them. For example, clustered bar charts rely on proximity to indicate grouping, so if elements are spaced too far apart, the viewer might not perceive them as related.

For example,  points that are close together in a scatter plot where will be interpreted as part of the same cluster and indicate a relationship or pattern.

• Closure : The mind tends to complete incomplete shapes or objects, filling in gaps to see a whole.

In visualizations like partial charts or visual elements with missing data, viewers will tend to "fill in the blanks." Designers can use this concept to simplify visual elements without overwhelming the audience.

For example,  a pie chart with a missing slice, will tend to force the viewer to interpret the missing slice as part of the  whole based on the available slices, even though the entire chart isn’t visible.

• Continuity : Elements that are aligned along a line or curve are perceived as related or part of a continuous flow.

Line graphs, flowcharts, or timelines use continuity to guide the viewer’s eye along a path, indicating the progression or flow of data. Lines or curves that smoothly connect points help viewers understand trends or patterns.

For example, a continuous line connecting data points in a line chart helps the viewer follow the trend, whereas broken or jagged lines might confuse the interpretation.

• Figure/Ground: People tend to separate objects from their background to focus on important information.

Contrast between the figure (foreground) and the ground (background) helps direct the viewer’s attention to the most important elements of the visualization. A clear distinction between data and background elements ensures that the viewer focuses on the relevant data.

For example, using contrasting colors for bars and a muted background in a bar chart ensures that the bars stand out, drawing the viewer’s attention to the data.

• Common Fate : Objects that are enclosed or bounded by a visible boundary are perceived as related.

Adding borders or backgrounds around certain elements can help create clear groupings or emphasize relationships. For instance, boxes around related charts or annotations highlight connections between them.

For example, using frames or background colors in data dashboards around different sections helps distinguish related information visually.

• Symmetry : Symmetrical objects are understood to belong together and are more pleasing to the eye.

Laying out the elements of a chart symmetrically creates balance and order, making visualizations more aesthetically pleasing and easier to understand. Symmetry also helps when designing dashboards or layouts.

For example, A symmetric layout in a dashboard with equal spacing between charts provides a balanced design that guides the viewer through the information without cognitive overload.

Nominal Data (Categorical):

• Consider using bar charts for displaying frequencies or counts of different categories.
• Pie charts can be suitable for showing the composition of a whole in terms of proportions.
• Use stacked bar charts or stacked area charts for comparing categories over multiple data points.

Ordinal Data (Ordered Categories):

• Bar charts or dot plots can effectively represent ordinal data, especially when the order matters.
• Ordered line charts can be used to show trends over time or another ordered dimension.

Interval Data (Numeric Data with Equal Intervals):

• Line charts are excellent for showing trends and changes in interval data over time or other ordered dimensions.
  
• Histograms or box plots can provide insights into the distribution and spread of data.
• For comparing values across categories, consider grouped bar charts.

Ratio Data (Numeric Data with a Meaningful Zero Point):

• Scatter plots are great for visualizing relationships and correlations between two continuous variables.
  
• ​​​​​​​Box plots and histograms can be used to display the distribution and summary statistics of ratio data.
• Line charts are useful for showing trends over time or ordered dimensions when you have ratio data.

Geospatial Data (Location-Based Data):

• Maps, including choropleth maps and point maps, are suitable for representing data that has a geographic component.
• ​​​​​​​Heatmaps can show the density of events or values in a geographic area.

Temporal Data (Time-Series Data):

• Line charts are a common choice for visualizing data over time, such as stock prices, weather trends, or website traffic.
• Time-series plots or Gantt charts can display events or processes over time.
• Multivariate Data (Multiple Variables):
• Scatterplot matrices allow you to visualize relationships between multiple variables in a grid of scatter plots.
  
• ​​​​​​​Parallel coordinates plots can show patterns and relationships between multiple variables.
• Bubble charts or 3D plots can represent data with three or more variables.

Hierarchical Data (Tree Structures):

• Tree maps or sunburst charts are suitable for displaying hierarchical data structures like organizational hierarchies or file directories.

​​​​​​​Network Data (Graph Structures):

• Network diagrams or graph visualizations are ideal for showing relationships between entities in a network, such as social networks or transportation networks.

​​​​​​​Text Data (Unstructured Data):

• Word clouds and textual heatmaps can visualize word frequency and importance in textual data.
• Sentiment analysis charts can represent sentiment scores over time or categories.