THE UX OF DATA Visualization Part 1
This past quarter at the University of Washington I took a class on Data Visualization. As a multifaceted Designer, I see the importance of Big Data in Future of Informing Design Decisions. Data is a powerful mechanism for telling stories. There will be a shift in how we communicate our Design stories.
To best understand our users, it is essential to collect data about user needs. In my Autumn quarter at the University of Washington, I learned a lifelong skill of Data Visualization. In this class, we learned how to tell stories through graphs, when and why we should use specific charts to display data. We worked with I.M.D.B data sets to tell stories about the movie industry. Also, for my project, I told the story of how products sold on Amazon.
I dove deep into the statistics, and I constructed Time-Series Analysis of product Sales Ranks over time. I made table lenses, maps, bar graphs, line graphs, and scatter plots. I performed distribution analysis, time-series, correlations, part-to-whole rankings, and multivariate analysis. I know now how to present data intentionally. Giving people access to their own data is a valuable gift. I could bring this new angle to your design team as you tackle challenging problems of today.
I am currently working on a small project that will help UX designers present data
What I Learned
What I read
Stephen Few Now you see it
What Story I told with Data
What are the Best Practices for Telling stories through data?
Forming key takeaways & Orient the Audience
When you present data, you always should think from the perspective of the viewer. What story is this graph trying to tell?
Questions you need to answer.
What does each axis mean?
What does one point on the graph represent?
What do the colors mean in the context of your graph?
Second you should address key takeaways on your graph
What type of graph is this?
What story is the data telling us?
Below are a few Guidelines we can stick to as UX designers as we present data
1. Time Series Analysis
What is shows
Trends, Variability, Rate of change, Co-variation, Cycle, Exceptions
Types of Graphs
Line graphs for Analyzing patterns and exceptions
Bar graphs for emphasizing and comparing individual values
Dot plots for analyzing irregular intervals
Radar graphs for comparing cycles
Heat-maps for Analyzing high volume cyclical patterns and exceptions
Box-Plots for analyzing distribution changes
Animated scatterplots for analyzing correlation changes
Best practices (pg.163)
Aggregating to various time intervals
Viewing time periods in context
Grouping related time intervals
Using running averages to enhance perception of high-level patterns
Omitting missing values from a display
Optimizing a graphs aspect ratio
Using logarithmic scales to compare rates of change
Overlapping time scale to compare cyclical patterns
Using cycle plots to examine trends and cycles together
Combining individual and cumulative values to compare actuals to a target
Shiting time to compare leading and lagging indicators
Stacking line graphs to compare multiple variables
Expressing time as a 0-100% compare asynchronous processes
2. Part to whole/Ranking Analysis (p.189)
Key takeaways data visualization should be …
Grouping categorical items in an ad hoc manner
Using Pareto Charts with percentile scales
Re-expressing values to solve quantitative scaling problems
Using line graphs to view ranking changes through time
3. Correlation (p.245)
How quantitative variables relate to and affect one another
Scatterplots for comparing two variables
Scatterplot Matrices for comparing multiple pairs of variables
Table Lens for comparing more than two variables simultaneously
Key Takeaways (p.247)
Best practices (p.266)
Optimizing aspect ratio and quantitative scales
Removing fill color to reduce overplotting
Comparing data to reference regions
Visually distinguishing data sets when they're divided into groups
Using trend lines to enhance perception of correlations shape, strength, and outliers
Using multiple trend lines to see categorical differences
Removing the rough to see the smooth more clearly
Using treillis and crosstab displays to reduce complexity and overplotting
Using grid lines to enhance comparison between scatterplots
Keeping intervals consistent
Selecting the best interval
Using measure that are resistant to outliers
Avoid using Pie Charts!
Use color to emphasis points on graph
Use size to emphasis points on graph
In sum, there is a lot we can learn from Stephen Few's Book Now You See It. I am currently still in the process of learning Tableu, and working on improving the way I tell my own stories with data. I truly believe that this is a skill we as designers should acquire. I've summarized the best practices of the most frequently used graphs, and briefly touched up ways we could improve how we tell our stories through data. As you read this, I am probably in process of creating something that will help UXer's learn of Data Visualization.