Projects
Resource Planning Analysis and Improvement @ CPI
As part of my Faculty data science fellowship, I worked with host company CPI to improve their manual processes and introduce data science methods in their business operations.
CPI works with a variety of clients, conducting technical development work to help them bring their products to market. As part of their work, they need to frequently estimate how long projects will take in order to allocate resources and inform staffing decisions.
However, CPI did not have a way of checking whether their predictions were accurate. I performed data engineering, combining data sources in order to assess the accuracy of their forecasting methods through time series analysis. I also created an automated tool that will help CPI check their prediction accuracy on a weekly basis.
Thanks to this project, CPI has been able to increase the accuracy of their predictions, and we estimate that, in the long run, it will save more than 162 staff hours per month by shortening the length of planning meetings.
Q-Suite @ Zebra Technologies
Q-Suite is a set of mobile applications for the enterprise, facilitating communication between employees and other work-related tasks. I was part of the team developing the user interface, written entirely in Dart and Flutter. Collaborating with the UX team, we created a responsive, robust and modern front-end
Accessibility was a focus point in this project, making the applications fully ADA compliant. I took responsibility for most accessibility-related tasks, including ensuring full compatibility with native accessibility services, auditing the applications to find any outstanding issues, and creating responsive apps that worked on devices of all sizes.
Flutter UI Component Library @ Zebra Technologies
A parallel project to Q-Suite, this consisted of developing a mobile-first user component library in Flutter, meant to increase consistency across apps and reduce development time. Aside from developing the components themselves, I also wrote most of the documentation, including best practices, and ensured the components were as accessible as possible.
LabelScan @ The University of Manchester
My Third Year Project at The University of Manchester was LabelScan, an Android app meant to make nutrition labels accessible to people with dyslexia and visual impairments.
Nutrition labels contain very important information about the foods we consume, especially for people with dietary restrictions, but their design leaves behind those with dyslexia and visual impairments. No nutrition label standard aims to make information accessible to them, and none leverage digital technology to enhance their accessibility. LabelScan attempted to bridge this gap, making the information of nutrition labels accessible.
Using Google’s ML Kit, the app scanned the nutrition label, extracted its text, parsed the information within and presented it back to the user in an accessible manner, presenting it visually rather than textually. Full compatibility with Talkback ensured the information was accessible to everyone.
You can check out the project’s source code here, and the written project report here.