Project detail

Data Generation Tool
Usability Test

CITIC Technology IP design

IP Design, Product Design

2023

CITIC Technology

About Synthetic Data Forge

Synthetic Data Forge is an online fake (Synthetic) data generation website targeting Data Scientists and Human Resource Officers. This Usability Test aimed to improve their beta version of the website.

Research Goal

  • Are users able to complete the tasks efficiently and without errors or confusion?

  • Which elements in the interface create resistance that mostly affects users unexperienced in using the tool?

Methods

This usability test includes 5 different methods, as listed below.

Jump to each section by clicking the names below.

User Interview

Goal

Test out How intuitive is the system for new users to navigate and complete essential tasks

Recrutement
  • Recruit 5 professionals in Data field

  • Participants handling sensitive governmental and industrial datasets

  • Participants lacked access to Reveal Global's production environment due to confidentiality constraints.

Method
  • used direct outreach and snowball sampling to recruit participants

  • semi-structured interview guide for data collection

  • Questions consists of open-ended questions focusing on the challenges in handling datasets and opinions on synthetic data and its applications.

Interview Script
Coded Interview Quotes
Key Findings
  • The data tools they commonly use have fixed formats that cannot be freely adjusted according to project needs. This rigid format makes it difficult for them to generate data that truly fits their work requirements

  • Manual data cleaning was one of the most time-consuming jobs for a data professional.

  • The tools lack the ability to annotate data fields or data logic, making it difficult for business and development teams to communicate the meaning of the data

  • The inability to capture real-world scenarios due to the lack of “unique datasets” available for those types of scenarios.

User Survey

Goal
  • Understand adoption, uses, issues, and trust of synthetic data.

  • Gather insights from HR professionals, government data scientists, and engineers through surveys.

Method
  • Created a draft of the survey questionnaire with over 20 questions

  • Seeking participants for feedback on the survey

  • Convenience sampling to gather participants with relevant knowledge and experience

  • Invited five individuals to participate in a pilot survey and provide feedback.

Survey Questions

Heuristic Evaluation

Goal

Identify usability issues that may affect user interaction and provide actionable recommendations for improvement.

Method
  • Team discussion to prioritize evaluation focus areas.

  • Identified key screens and flows based on their importance to user experience.

  • Used Nielsen’s 10 Usability Heuristics as evaluation criteria.

  • Severity Assessment Based on impact on user experience and task completion

Analyzation Script

Nielsen’s set of heuristics (Nielsen, 1994) that we used for our heuristic evaluations

Team members' first note-taking

Most Risky Issues

Key Findings & Recommendations
  • Finding: System status lack of color contrast, causing low readability.

  • Suggest: Modifying the font color of the header or the contrast. Develop Structured layouts, icons, and visual cues

  • Finding: Inconsistent UI elements, such as "logout" button appears in different locations in different pages.

  • Suggest: Bring Log-out button to the top right. Develop standardization UI components.

  • Finding: Absence of error prevention and recovery notification when user types in formula variable & invalid input numbers.

  • Suggest: Add contextual help features, like question mark icons or an in-line help sidebar, Improving error recovery through explicit error messages and inline notifications

  • Finding: No contextual assistance or intuitive documentation, causing difficulty in understand features.

  • Suggest: Including a preview function with example real-world data examples


Comparative Evaluation

Goal

Identify how Synthetic Data Forge compares with competitor products for it's target users and RGC in the following respects:

  • Desirability (customer needs and market demand)

  • Feasibility (technical and operational capability)

  • Visability (business sustainability and economic value)

Method

Categorizing competitors based on their relevance to our product, considering:

Direct Competitors

Products providing the same primary function and targeting the same audience. 

Inirect Competitors

Products offering similar functions but aimed at different user groups or purposes.

Partial Competitors

Products sharing some core features but differing in scope.

Parallel Competitors

Products in different domains but with comparable interaction styles, workflows, or user experience patterns.

Analogous Competitors

Products or services in entirely different domains, but similar in operational style or interaction approach.

Criteria
  1. What core features and capabilities does the product offer?

  2. Who is the primary user or customer of the product?

  3. How do users typically engage or interact with the product?

  4. Can the product handle growth and high data volumes?

  5.  How is the product priced and what monetization strategies are used?

  6.  What level of product adaptation is available?

  7. How closely does the product generate synthetic data mimicking real-world data?

Comparative Evaluation Charts & Findings

Usability Test

Goal

Identify how Synthetic Data Forge compares with competitor products for it's target users and RGC in the following respects:

  • Desirability (customer needs and market demand)

  • Feasibility (technical and operational capability)

  • Visability (business sustainability and economic value)

Method
  • Conducted 1 usability pilot test and 5 usability tests

  • Participants had experience working in data field, but never used Synthetic Data Forge before.

  • Each test included a pre-test survey, 6 tasks that focused on the search, and wrapped up with a debrief interview.

Participants

  • Work in data field (Data scientist / Data analyzer / Data student)

  • Never used Synthetic Data Forge before

Pre-test Survey

  • Consists of 8 questions

  • Asking about demographic information of the participants.

Tasks

  • Focusing on the commonly used features of the Synthetic Data Forge website.

  • Record think-out loud thinking process

  • Calculate time spent on each task & subtasks

Debrief Interview

  • Consists of 8 questions

  • About participants' thoughts on usability, visual design, experience and general understanding of the site.

  • Coded as documentation for analyzation

User Testing Recording
Scripts & guides during the user test (including Tasks, Task Instruction File, and User Content Form
Data Log-in Form (Note system to capture, measure, and analyze users' every small moves)
Performance Data Form

Unclear navigation and layout

  • Finding: Participants found the navigation confusing, often struggling to locate key pages like instructions or download, and had difficulty moving smoothly through the workflow.

  • Suggestion: Restructure the navigation to reflect the linear data generation workflow, add breadcrumbs or a stepper component to visually indicate user progress

Lack of Visual Feedback and Confirmation

  • Finding: Users lacked clear feedback after saving or uploading actions, causing confusion and repeated manual input.

  • Suggestion: Introduce clear, in-line success messages or toast notifications. Provide visual affordances such as progress bars or icons that confirm a completed step. 

Difficulties in Correcting or Editing Inputs

  • Finding: Users found it difficult to correct or edit inputs, as the interface lacked clear options for going back or safely adjusting previous entries without losing progress.

  • Suggestion: Add a persistent “Edit” or “Back” button in each step, add auto-save functionality or clearer “save” states

Minimal Instructional Support for New Users

  • Finding: New users lacked orientation upon login and had difficulty mapping tasks to system features.

  • Suggestion: Create a welcome dashboard that provides a brief overview and quick links to key sections. Offer optional onboarding, Add hover-based tooltips to clarify icon meanings and terminology

Visual Design is Functional but Flat

  • Finding: Users appreciated the platform’s simplicity but found it visually bland, making it harder to focus on key areas or actions.

  • Suggestion: Incorporate a more vibrant and accessible color palette, emphasize primary actions

Mettalic shape background image

Contact

Let's Get in Touch

Let's connect and start with your project ASAP.

Or Call 917-521-9252

Mettalic shape background image

Contact

Let's Get in Touch

Let's connect and start with your project ASAP.

Or Call 917-521-9252

Mettalic shape background image

Contact

Let's Get in Touch

Let's connect and start with your project ASAP.

Or Call 917-521-9252