By – Surabhi Gore
( User Experience Lead | Digitate)
By – Dr. Maitreya Natu
(AI Scientist | Digitate)
Excellent analysis, if not presented well, becomes practically unusable. The only way to industrialize the power of analytics is to pair it with equally effective visualization. Effective visualization brings explainability to the analytics findings. It helps the users understand the intuition and the thought processes that went into deriving the findings.
If designed carefully, visualization can ease the task of inferring a complex analysis. In this document, we present three case studies to demonstrate the power of visual analytics. We have chosen these case studies to highlight how large, complex data can be rendered through effective metaphors and how insights can be derived with ease.
Case study #1: Uber rides
Uber uses data-dense visualizations to analyze commutes across a region based on daily, weekly, and monthly statistics. It also performs real-time analysis of pick-and-drops and customer trips to understand demand, traffic patterns, inefficiencies, and so on. These visualizations rely solely on the raw, real-time data flowing through their network1.
Maps are usually the best metaphor for displaying geographical data. Uber has used maps in creative ways by developing interesting layers on top of maps to overlay various statistical observations on maps and adapt to real-time feeds. Consider an example of rendering vertical bars on the geographical maps where the height of a location indicates the number of rides picked-up from that location. This visualization helps find pick-up hot-spots, morning and evening patterns, among others
Case Study #2: Pandemic statistics
Fighting a pandemic requires continuous monitoring of the infections across the world to plan and take corrective actions. Visual Capitalist2 analyzes the growth trajectories of the number of cases across days and across regions. Such use cases require the utmost adaptations possible. Furthermore, this visualization needs to get updated on a timely basis as case numbers are published.
There is no need of fancy visualization when a simple line chart can do the trick. Line charts are very effective in showing change over time . Chart can be made effective by supporting different forms of filters and aggregation, and the ability to zoom-in and zoom-out. The shape of the curve itself speaks for trends, outliers, and changes in behavior. Visual Capitalist very effectively uses line-charts to plot number of cases across regions. It uses the trends and patterns of these curves to derive insights about a region’s success in fighting the pandemic.
Case Study #3: Wind map
Another example where visualization plays a very effective role is in identifying the site for a wind farm. The velocity and direction of wind plays a key role in not just selecting a site for the wind farm but also selecting suitable wind turbines and optimum installation height to maximize the output. Ventusky4 very effectively uses wind maps3 for this purpose.
Speed is represented by lines moving slow or fast, and the direction is represented by which way the lines are moving . Ventusky4 overlays this with additional data such as cloudiness, temperature and air pressure. Such visualizations prove to be a strong tool where decision making is done based on atmospheric factors.
Visualization by Design
With increasing digitization, all industries are trying to tap into the power that data analytics can unleash. However, the lack of explainability directly impacts the confidence and the rate of adoption of complex analytical findings. The only way to industrialize the power of analytics is to pair it with equally effective visualization. It helps the users understand the intuition behind the analytics and focus on the areas that need attention.
Enterprise data presents several interesting visualization opportunities to render historical behavior of metrics, correlations of events, and predictions of processes, among others. The visualization also has an element of offline versus real-time rendering of information. ignio uses various metaphors to best render this variety of information. One such metaphor is the Gantt widget.
Gantt chart is a perfect metaphor to render space and time dynamics of a process. However, this widget needs various enhancements to ensure that the right amount of information is presented and that the user’s attention is brought to the right areas that need attention. This is done by enhancing the basic Gantt widget with various aspects such as folding and unfolding of the Gantt in the dimensions of both space and time, by highlighting critical paths, by color-coding different sections, and by presenting a comparison of multiple Gantts. Similarly, there are several other visualization examples in ignio that demonstrate the importance of rendering the data to make the analytics easy to comprehend, and to systematically guide the user to the areas that need attention.
