Booster Framework is an application that allows users to visualize datasets using a collection of advanced data visualizations, built around machine learning and AI algorithms.
Could we provide smaller parties with data visualization and forecasting tools in web-based product?
Our developers at MAQ Software had a good collection of algorithms they've written for forecasting and analytics, typically more advanced than what you would find in typical business intelligence software like Power BI and Tableau. We had an opportunity here: we do a lot of great data analytics work, which we use for clients all the time. Could we provide smaller parties with data visualization and forecasting tools in web-based product?
Originally this application was built using a template without any input from the design team, and it was only used internally. However, this original tool was disorganized and frankly not commercial, so the project manager brought the problem to me.
Mr. Project Manager expressed a wish to bring this tool to a broader commercial audience and make it more user-friendly. He also mentioned wanting to expand the functionality, so users could possibly pay for different subscription amounts to house different amounts of data and to make the tool scalable for larger organizations.
Make the complex easy
The tool should guide the user seamlessly without much explanation. Additionally, the tool should also incorporate more teaching moments, especially for uncommon data representations.
Give it room to breathe
Space should be redistributed to give enough room for what's important, especially since the user will need to work with complex data that can make the design feel cramped.
Make it customer-ready
The tool already existed using a web template, but this had no branding, enticing visual elements, or features for ease of use.
Stretch goal: Allow for multiple use cases
The tool should allow for expansion and more features. Eventually, admins would be able to control data usage and manage peers, while ordinary users could be restricted to specific parts of the tool.
Before remodeling the Booster Framework, I gathered some research about potential end users and competitor products, such as Power BI and Tableau.
The target user is someone who has some familiarity with data, but perhaps is not an expert and so will need explanations for complex data visualizations. According to project managers and stakeholders, they wanted to target the application towards small companies or people who would use this for business.
Based on this information, though, the structure of the old tool was way too complicated for the target audience.
Here's an example of the journey of a user who has already has an account, but wants to explore a new dataset:
The sign-in process should be straightforward and familiar.
The user uploads either an Excel file (.xslx) or a CSV. These datasets may be complex with many rows.
The user sees a summarized version of their dataset, and performs basic data cleaning such as renaming columns.
This is the "visualization" portion of the tool. The user selects a visualization, then selects the metrics they would like to view and edits any parameters needed for the visual.
Simple, right? But the original tool had way too many choices upfront, making it difficult for users to find what they need to progress and follow the ideal flow. There were nine items in the navigation, and half of those items were not self-explanatory.
I remade the user flow so it's more logical and easier for the user to navigate:
While iterating over possible solutions, there were definitely challenges in simplifying a complex application without making it too simple or confusing. In addition, the Booster Framework has a lot of potential, so it was important not to go overboard with features that either are not helpful for the user to reach their goal, or that would be to resource-heavy for the stakeholders and developer team.
This tool can't (and shouldn't) do it all
Especially compared to possible competitors. The application doesn't accept all data types, and will lack the formatting options available from competitors.
It's a complex topic for me too
While I was familiar with many representations for data, some of the more complex visualizations and transformations needed more study.
A couple of insights from working on the design:
Eventually we were able to create an application that was easier to use for those who aren't as familiar with complex data visualizations. Feedback on the application was very positive, especially the look and feel. Project managers started to see that our design teams were capable of making things usable and aesthetic at the same time.
The Booster Framework, didn't reach its stretch goals, however. Due to decisions from upper management, we ended up not selling this as a product on its own. While we did end up also removing some of the "commercial" aspects, like the fancy landing page, we did keep the organizational structure I proposed and continued to use the tool internally or when presenting to clients.
It elevated our expectations for future designs.
Project managers were impressed with the look and feel of the product. Wanted to hold onto this design since it was so good, even though they weren't fully able to manifest this.
Design teams starting being involved earlier.
The tool already existed using a web template, but this had no branding or enticing visual elements.
Well first, I wish the name was different, at least more descriptive of the tool. But overall, I loved this project since I was able to take complete control of the design process, and create something I think turned out pretty well, even if we didn't ship it in its entirety. If I could, I would definitely go back and tell myself to optimize the application for better accessibility, but overall I think the project was a success.
