Company and Sector: Verta Inc (AI/ML Platform)
Objective: Design the Search feature for the Verta application
Role: Design lead
Project duration: 3-4 weeks
Overview
Verta is an application used to track data models, experiments, and other aspects of governance.
Needed a Search feature to locate elements belonging to various categories like Datasets, Experiments, Projects, or People.
Also needed to surface metadata about each search result.
Design Process
I broke down the design solution into three parts:
1. Entering the query in a Search box, features surfaced in the context of the Search box.
2. Displaying the Search Results
3. Conceptualizing an advanced Search feature which allowed more controls over crafting the Search query
Final result:
The simple Search query approach was chosen as the first version to develop, with the advanced version tabled for another time. The Search result display was built as proposed and made searching for assets infinitely easier.
The Search design consisted of 2 parts:
1. The Search interaction where the user enters a Search query
2. Displaying the results of the Search query
Search query:
I proposed two approaches to designing the Search query. One was a simple text typing and the system would provide suggestive terms in a drop down, and another was an advanced version targeted at data scientists and power users of the app. This advanced version would provide multiple fields to the user to choose from as well as a choice of operators (and/or/not) from a drop down to build a complex search query.
I also proposed ways of surfacing frequent queries and a way to mark them as favorites for a certain user. A simple display of a '*' next to a query provided the user to mark a query as their favorite.
Search result display:
The search results were displayed as cards with the matching attributes highlighted. The search term would yield various types of results (e.g. datasets, experiments, projects etc) and I proposed to use iconography to differentiate the different types of results.
Final result:
The simple Search query approach was chosen as the first version to develop, with the advanced version tabled for another time. The Search result display was built as proposed and made searching for assets infinitely easier.
Advanced Search
Explore the ability to save and retrieve frequently used search queries
Explore advanced search features for power users to pinpoint an exact experiment amongst thousands.
Learning: As the lead designer working on creating design components for the first time, the level of details needed was unclear. So my first pass on the designs lacked a lot of details.