Holistic User Experience for
Big Data Analytics

When I get to design big data analytics platform, I often worked with data scientists, data analysts, and industry researchers. Often times in my experiences of designing enterprise big data analytics platform, data scientists have been working on the certain specification and algorithm for bid data with specific business goal in mind. Designers are often brought in to advocate the usability and enhances the experiences of human interaction with big data.

Challenge

Big data platforms can get very complex and have many layers of information. If we don’t pay attention to the human interaction side of the usability, dealing with big data can get none usable very quickly.

Starting point

Most of the time, there were interactive prototypes that generated data produced by data scientists and engineers. This type of working prototype is usually to prove that the platform spits out the data and displays them correctly. I as a designer, this has been the usual starting point when I design enterprise level of big data analytics.

 

The following image is the sample of prototype to show how the UX approach will be applied as an example.

BigDataAnalytics-example-screens-01

After receiving the prototype, always the first thing to do is to figure out the problem. Problems are sometimes not clear. But it is the most important thing to do, identify the problem. Without this part, there are always some possibilities that the design will not be relevant for the product. In addition, it is not possible to solve all the problems at once. Therefore, clarifying the problems will help to prioritize what to solve.

Process to clarify problems: Example 1
BigDataAnalytics-example-screens-components

In order to understand the problems, I started to put color on each component in certain color based on their functionality. The intention here is to see if there are any patterns and rules in the interface, which components are functioning in similar way, what kind of information are displayed, and so on.

 

After this process, several problems were identified. Some of the example problems are the followings:

– Most of the interaction processes were put into one screen. This was creating the interface to overcrowd with information.

– Data visualization sometimes displayed using only partial screen. It made the interaction activities to the data difficult.

– There were action buttons to perform different executions placed in various places.

– There were inconsistencies in the UI usage.

– Because of the reasons above, information hierarchy was not very clear for a user.

Process to clarify problems: Example 2

There were other cases when the prototypes had similar problems to the example 1, but the screen jumps to different screens, and it added another layers of unfriendly user experience.

In this case, we went back to mapping technique to clarify some flow before we start the screen level of user experience.

BigDataAnalytics-siteMap-perspective
Finding the UI component pattern

Blocking each components of the prototype and mapping the whole existing system flow made it more clear for us to see the UI pattern. The following list is just a simple example of component lists.

– Menu

– Data

– Data table

– Check list

– Radio button list

– Execution button

 

 

Listing things up helps to see what information should take how much real estate of the screen for interaction purposes as well before putting onto the design.

Restructuring the information flow with consideration of user interaction

From the above analysis, there was a certain interaction flow between a user and a system became more clear. This flow is a very simplified version as an example, but I hope this will help to demonstrate some UX approach idea. I often think of this interaction as like a conversation flow between two entities.

 

1. User enters what data to show.

2. System generates and displays the data.

3. User analyzes the data, interacts with the data.

4. User decides and take action based on the data.

5. System executes the action and gives response to a user.

6. User will finish the session or go back to step1.

 

After witnessing these steps, it is clear that there is always some user inputs to generate the data, and the execution for the whole process. So keeping that data input interaction area consistent will provide the cohesive experience for a user. Then, providing some other contents and data in the main area to organize the whole information architecture.

BigDataAnalytics-example-screens2-02
Learning

For this type of design work, the product stage hasn’t developed enough yet and the end user is not defined clearly either. Designers have to make a best guess for the usability. This can be challenging because it feels like designing for the unknown users because we usually stay really focus on the user profile. My learning was to broaden user profile means consider more natural human factors and usability point of view from none technical user point of view actually helps a lot to tackle the design. For the interface side, it’s usually too complex and break it down the information architecture and identify every elements that are causing the complication helped a lot.

 

Again, this interface is not designed with the consideration of an end user, therefore it does have a lot of limitation. In addition, it is not considering any factor about responsive or different medium usage, so there are a lot more work to be done. However, applying some cohesiveness to the flow is like building a bone structure of the information flow. This approach clarifies the possibility and limitation in the future requirements and it will definitely help to scale up from here.