OVERVIEW

Our capstone project team was commissioned by the CEO of Stardog, a fast-growing Enterprise Knowledge Graph Platform provider, to redesign their existing knowledge graph product that could be used not only by technical users but also business analysts who typically rely on IT to create a graph for them. We identified a need for a simpler data unification process. We collaborated closely with their product and engineering team on a mission to design various Graphical User Interfaces (GUIs) that would allow non-technical users to do data mapping, data modeling on their own without any coding.

I managed a team of 6 designers and collaborated with stakeholders and customers, led the design operations, user research and product evaluation, conceptualized solutions, and owned 4 critical features in the final interface design.

CLIENT

DURATION

Aug 2019 - May 2020 (9 months)

ROLE

Design Operations
User Research
Usability Testing
Product Design

TEAM

Jashan Gupta
Modassir Iqbal
Yirang Choe
Monikka Ravichandram
Aravind Jembu Rajkumar
Natalie Yeh

TOOLS USED

PROBLEM STATEMENT

Despite its benefits, the current process of creating knowledge graphs requires technical expertise and is time-consuming. Stardog realized the need for a simpler alternative to expand its user base to non-technical people.

The goal of this project is to design a tool that helps users with limited technological skills to do data mapping and data modeling.

BACKGROUND

Understanding knowledge graphs and different user segments.

What is the Knowledge Graph?

A knowledge graph (KG) is a flexible graphical representation of data derived from a variety of structured, semi-structured, and unstructured data sources that are connected together to create machine-understandable knowledge.

Enterprise data in large companies is stored in silos and fails to provide insights required to make better decisions. KGs allows users to build flexible models and connect data from different silos.

The process of creating and using KGs (detailed workflow) requires a number of large, complex steps which are illustrated below. Our project focus on frame 2 to 4, supporting users to do data modeling and mainly data mapping.

Storyboard illustrating the knowledge graph creation process.

User segments

RESEARCH METHODS

Understanding the user's mental models and exploring existing solutions.

SME Interviews

Knowledge transfer sessions with Stardog engineers.

Competitor Analysis

Analyzed existing solutions in knowledge graph space.

Structured interviews

Conducted interviews with 8 High-techies and 3 Low-techies.

Contextual Inquiries

Conducted 5 interviews to capture observational data.

Usability Testing

Evaluated lo-fi prototype with 5 participants.

Affinity map containing detailed findings. Explore Affinity Map on Miro

Current Stardog Customer Workflow

Usability Testing Notes

FINDINGS

Core user needs

1. Modeling and mapping are very closely tied.

When users work on mapping, they need to edit/add new nodes or properties as needs arise. So we treated each project as a virtual graph where users can create data models and do the mapping on the same canvas.

4. A subset of data is selected before users start mapping.

Users select some data-columns to work with before mapping so they don’t get overwhelmed by a huge amount of data. So we created a data-field selection user flow.

2. High-techies prefer coding.

High-techies prefer coding because they are more comfortable with writing code than using a visual interface. So we decided to provide the flexibility to switch between synced visual and code modes.

5. Properties are defined separately.

Users may define properties and classes separately and later assign the properties to the classes and map them. So we created a new place for defining properties separately.

3. Auto-mapping is the default starting point for techies.

The majority of current users use Auto-mapping functionality as a starting point. So we highlighted the ‘Auto mapping’ feature when users first create the virtual graph project.

6. 'Click to map' vs 'drag and drop'  to map.

Users found both 'click to map' and 'drag and drop' mapping functionalities to be useful. So we supported both 'click to map' and 'drag and drop to map' interactions.

USABILITY TESTING

Evaluating lo-fi designs and iterating over it

1. Don't force users to name their project when they create a new one.

2. Reduce the efforts of the data selection and mapping process.

3. Created a flexible way to add properties and support different use cases.

FINAL DESIGNS

The code-free knowledge graph creation interface

1. Virtual graph + Automapping flow

2. Data Modeling Flow

3. Data Mapping Flow.

LEARNINGS & TAKEAWAYS

My major learnings from the project

Stardog was definitely the most challenging product that I worked on considering its unique industry space targeted towards developers. The unique challenges posed by this project forced me to push my limits and made me more comfortable managing both the research and design roles.

Leading a team of 6 designers made me think from a perspective of design operations and helped me realize the importance of different perspectives while designing something. Throughout the project, I got to lead Design Operations, Research, and Design which helped me grow tremendously as a designer.

Reflecting back, most of the challenges occurred because of our limited knowledge of this space, and I would have spent more time on Desk Research and learning Modeling and Mapping through code so that I could have understood the complex use cases better.