Information Visualization
Information visualization is rarely about visualizing fancy charts and is more about conveying the underlying meaning of a dataset. It is about telling a compelling story behind the data and facilitate human cognition.
Understanding data and information
Just like any other user experience work, I believe information visualization also requires structured design thinking. I have been fortunate to get involved in a couple of visualizations projects that have helped me understand and experience the nuances of and appreciate the process that forms its foundation.
I summarize here what I see as design thinking for visualizing the data story.
Identify the problem
As always, it is crucial to identify
- who is the user?
- what problems are we trying to solve?
- what is the user trying to accomplish?
- what are their needs?
- what are their challenges?
- why is the current system not solving their needs?
- where are the gaps and how would visualization bridge the gap?
It is imperative that we get to the grass root level of knowing the users and their context so as to identify and solution the right problem.
Visually laying out the persona's thought process, as below, is helpful to empathise the volume of information required and to better understand the user needs.
Define the goal
Where do we want to reach, what is our goal? Each of the thought in the thought process was broken into questions and potential answers that may pin point to individual user needs. Parallel usability testing and user interviews were carried to verify that the assumptions taken towards the potential solution are correct.
Understand the data
Know what type of data do we have in place to start mapping it to the goals. In order to do that, we will need to execute the following steps:
Defining the data
For each of the questions above, there was a potential answer. But how do we map these answers to visualizations? To answer this, we should first know what type of data will be used?
- Will it be a quantitative data that has a numerical value?
- Will it be an ordinal data that has underlying time order such as weeks and days?
- Will it be a categorical data that is discrete such as names?
Data dimensions
Data can have multiple dimensions or attributes. For example, a customer might have multiple dimensions such as first name, last name, address, contact etc.
- Univariate data: A single variable is studied at a time. For example, speed, weight etc. Important inferences using mean, spread etc can be obtained using univariate data. Univariate data are simple and can be used for simple interactions such as filtering of big data sets.
- Bi variate data: As the term suggests, there are two dependent variables are studied. Sale versus time of the year is a good example. Comparisons and relationships are simple usages of bivariate data. Basic level visualizations stops at this level of data dimension.
- Multivariate data: These are more complex because of more number of dependent variables. More the number of dimensions, more confusing it can get to comprehend.
Data structure
- Linear: data are linear and can be shown in tables and vectors etc.
- Temporal: data changes over time
- Spatial: data relates to real world showing geographical relationships (eg: office floor plan or count of something per country
- Hierarchical: such as office positions
- Network: such as social connections
Interactions
Finally, it is time to pick the visualizations. Visualizations could be static, tranformable or manipulative depending on the need.
Below is a sample screen showing the IP Status data and actions that are due for the Attorney persona to act upon. Due to NDA the rest of the design can not be shared at this moment.