In this project, we’re building a graphical user interface for non-technical people who have limited knowledge in coding to do data modeling and data mapping. The current process of data modeling, mapping, and querying can only be done by writing code. However, to expand the customer base, we designed a tool that could bridge the gap between technical users and non-tech users so that both parties can easily work on projects and maintain the knowledge through the tool.
This is the capstone project worked by the UMD team partnered with the company, Stardog, a leading knowledge graph platform for enterprises including NASA, CapitalOne, Dow Jones, and Bank of America.
Sep 2019 - May 2020 (8 months)
Low-fi & Hi-fi prototypes
Aravind Jembu Rajkumar
Knowledge Graph building process includes collaboration between different teams, including high-techies and low techies. Design a visual tool that could be used by people with different levels of expertise in coding to expand the customer base.
The current Knowledge Graph building process is sufficient for engineers who are experienced in writing code. However, it’s too difficult for non-technical users, such as SMEs or business analysts, who would be helping to design and build a knowledge graph. How might we design a tool that helps non-technical users to engage more in the process?
Knowledge Graph is a data management tool that connects all data, data sources, and databases of every type. Enterprise data is usually stored in different data silos, which failed to provide insights because they’re disconnected. Building a knowledge graph help to connect all enterprise data and answer questions.
source: stardog website. https://www.stardog.com/how-it-works/
Our client is very satisfied with the final deliverables. Although there are some advanced use cases not considered due to project scope, they’re using our design as the starting point to build the visual editor, and we can’t wait to see the result!
Below is the prototype walkthrough. It contains mainly 3 parts, automapping, data modeling and data mapping.
USER RESEARCH & FINDINGS
Our project focus on frame 2 to 4, supporting users to do data modeling and mainly data mapping.
1. Modeling and mapping are very closely tied. When users work on mapping, usually they need to edit/ add new nodes or properties as needs arise.
2. High-techies prefer coding. They are more comfortable with writing codes rather than using a graphical interface.
3. Auto-mapping is the default starting point for techies. The system auto-generates the data model and mapping for them which is editable.
4. Users want to select some data before they start mapping. They select a few fields that they need to work on before mapping.
5. Defining properties separately. We discovered that users often define properties and classes separately, and later assign the properties to the classes and map them.
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.
KEY DESIGN DECISIONS
Use Blueprint.js as a starting point and create the icons, colors, components and states for different UI.