Oxford Sentence Annotator:
Empowering startups with
efficient text annotation
Product made by research team of
Introduction
The Oxford Sentence Annotator is a web tool developed in collaboration with Oxford University's research team to accelerate the annotation process for text corpora. The tool was specifically designed for the Startup Accelerator Programme at Oxford University, catering to startups dealing with extensive databases and the need to train machine learning models.
Annotator's project screen from Oxford Sentence Annotator
Challenge
Startups in the Accelerator Programme required efficient text annotation tools to train their machine learning models effectively. The challenge was to create a user-friendly, intuitive, and efficient web-based annotation platform catering to three key user roles: administrators, reviewers, and annotators. This tool needed to streamline the annotation process while maintaining a high degree of accuracy and consistency.
Design process
Solution
Collaborating closely with the Oxford University research team, our cross-functional team, consisting of a product manager, product designer (myself), engineer, machine learning specialist, and customer development specialist, embarked on the creation of the Oxford Sentence Annotator.
The process began with scoping, defining goals, and specifying requirements. I created detailed user flow charts for the three user roles, followed by wireframes that outlined each step of the scenarios. I then translated these wireframes into high-fidelity clickable prototypes, enabling the customer development specialist to gather valuable feedback.
The tool offers a dynamic and user-friendly platform that allows administrators to create and manage projects seamlessly. Reviewers can efficiently address disagreements in annotations, while annotators benefit from an intuitive interface and keyboard shortcuts for rapid annotation.
User flow
Wireframes
My role
As the product designer, I was responsible for shaping the user experience and user interface of the Oxford Sentence Annotator. I worked closely with the research team to understand the specific requirements of administrators, reviewers, and annotators. My role included designing the login page, project overview page, annotation page, review page, statistics page, and other critical components of the tool. I created the user flow charts, wireframes, and high-fidelity prototypes, ensuring a seamless and intuitive process for administrators, reviewers, and annotators. I collaborated closely with the team, proactively identifying issues and providing solutions to enhance the product's design and functionality. I ensured that the design facilitated a smooth workflow, minimal distractions, and clear feedback for users.
List of projects for annotation
Reviewing process
Editing project by admin
Project statistics
Results
The pre-implementation phase, spanning two months, saw the successful creation of a comprehensive design solution. Feedback from the customer development specialist and the research team contributed to iterative improvements in the design. As the engineer began the implementation phase, I supported the process by creating a style guide in Figma and handoff in Zeplin. Additionally, I wrote an article in Confluence to document implementation rules.
The Project Manager provided feedback on my work, stating, "It was great to work with Evgen. The communication was clear and prompt, and he professionally delivered a great design in a short time. He proactively identified issues and provided solutions for our product design, and worked well within our team."
Style guide
Conclusion
The collaboration with Oxford University's research team resulted in the successful design of the Oxford Sentence Annotator, catering to the needs of the Startup Accelerator Programme. The user-centered design approach, collaboration with the team, and iterative feedback loop contributed to creating an efficient and user-friendly annotation tool. The design process effectively addressed the challenge of streamlining the annotation process, thus aiding startups in their machine learning endeavors within the Startup Accelerator Programme.
Thanks for watching!
