Memorise
Overview
Ai assisted mnemonic generation iOS app for students
Responsibilities
UX Researcher
Full Stack Developer
Designer
UX Researcher
Full Stack Developer
Designer
Context
January 2024 - April 2024
Bsc Computer Science Final Year Project
Supervised by Pieter Joubert
Tools Used
Open AI
Figma
Canva
Python
Flutter
Visual Studio Code
Firebase
Git
Gitlab
Design Brief
As a degree apprentice student working for PwC, for my final year project we had to ensure our project aligned with a business need.
As a degree apprentice student working for PwC, for my final year project we had to ensure our project aligned with a business need.
Problem Statement
PwC employees often face the challenge of completing demanding professional qualifications, such as the ACA (Associate Chartered Accountant). Many professionals struggle with memorising large volumes of information, which can lead to burnout and inefficiency. This highlighted the need for an innovative solution to help them efficiently retain complex content, as traditional memorisation methods often proved inadequate and required more modern, personalised techniques.
Methodology
As the project was to be submitted as an academic dissertation, I decided to blend human-centred design practices with Agile methodology for development, supported by a well-researched proposal as the foundation for my overall methodology.
As the project was to be submitted as an academic dissertation, I decided to blend human-centred design practices with Agile methodology for development, supported by a well-researched proposal as the foundation for my overall methodology.
(1). Research
Academic Research - Key Findings
To ensure the integrity of my project and to learn from the leading experts in memory, I conducted an extensive literature review into memory, Artificial Intelligence and existing memorisation solutions for revision.
Insights
Dual encoding of a memory using both visual and written is more effective
Personalising memorisation content to the user can lead to more optimal learning
Nothing in the market which provided Auto generated mnemonics for revision
User Resarch
To learn more about taking professional qualifications whilst also working a full time, I opted to conduct semi-organised interviews with PwC employees to learn about their pain points.
Insights
Memorisation of content was a significant challenge
Wanted a revision tool that was quick to use and required little effort to learn
Users wanted accurate specific content tailored to their qualifications syllabus
(2). Design
Persona Creation
To encapsulate my user research I decided to create a persona in order to be able to refer back to when creating my User Stories and system requirements
Wireframing
Using Figma, I started with low-fidelity wireframes to ensure a solid user experience (UX) for the application. From there, I created high-fidelity wireframes to explore layout options. The design emphasises simplicity and ease of use, featuring a soft colour scheme—light beige backgrounds with dark grey accents—to minimise cognitive overload.
Using Figma, I started with low-fidelity wireframes to ensure a solid user experience (UX) for the application. From there, I created high-fidelity wireframes to explore layout options. The design emphasises simplicity and ease of use, featuring a soft colour scheme—light beige backgrounds with dark grey accents—to minimise cognitive overload.
(3). Implementation
Technology Stack
The front end was developed using Flutter, while Firebase Firestore was used as the backend database for storing user data. OpenAI was integrated for generating mnemonics, and GitLab was used for continuous integration.
AI Powered Mnemonics
The app used the OpenAI API to generate personalised mnemonics based on user input, processed through a cloud function to ensure consistent results.
Iteration
By creating MVPs and gathering feedback from stakeholders and users, I was able to streamline the UX by removing an unnecessary step in the mnemonic generation process. Users reported that this optimisation made the process feel faster and more intuitive, making it easier to create mnemonics with minimal effort.
Authentication
Profile Creation - Add User Intrests
Mnemonic Generation
AI Generated Mnemonics
Accessing Created Mnemonics
Personalising Content to Syllabus
(4). User Testing
User Study
I conducted a user study with 15 participants to evaluate the effectiveness of my application compared to two leading competitors, Anki and Quizlet. The study assessed how well each application helped users memorise sections of the ACA qualification.
Participants completed an immediate recall test, followed by a second test one week later. Students using Memorise scored 25% higher than those using Quizlet, which ranked second, in the immediate recall test and 35% higher in the test one week later.
System Usability
Using the System Usability Scale, I got the participants of the user study to analyse the applications usability. Memorise scored 87.03 putting it into the excellent category.
Reflections and Acknowledgments
Initially, I developed a different solution for this project: a memory palace creator environment built in Unity. However, after consulting with stakeholders, I decided to pivot and realign my solution with the needs of my target audience. This was a crucial lesson in understanding the importance of designing for users. Following the pivot, I had a much shorter development timeline to learn mobile development and build a functioning application. As a result, I worked long hours every day up to the deadline to complete the project. Though demanding, it was an incredibly rewarding experience.
I would like to thank my supervisor, Professor Pieter Joubert, for his support and guidance throughout this project. He was always available to discuss my work and provide valuable advice.
Memorise
Overview
Ai assisted mnemonic generation iOS app for students
Responsibilities
UX Researcher
Full Stack Developer
Designer
Context
January 2024 - April 2024
Bsc Computer Science Final Year Project
Supervised by Pieter Joubert
Tools Used
Open AI
Figma
Canva
Python
Flutter
Visual Studio Code
Firebase
Git
Gitlab
Design Brief
As a degree apprentice student working for PwC, for my final year project we had to ensure our project aligned with a business need.
Problem Statement
PwC employees often face the challenge of completing demanding professional qualifications, such as the ACA (Associate Chartered Accountant). Many professionals struggle with memorising large volumes of information, which can lead to burnout and inefficiency. This highlighted the need for an innovative solution to help them efficiently retain complex content, as traditional memorisation methods often proved inadequate and required more modern, personalised techniques.
Methodology
As the project was to be submitted as an academic dissertation, I decided to blend human-centred design practices with Agile methodology for development, supported by a well-researched proposal as the foundation for my overall methodology.
(1). Research
Academic Research - Key Findings
To ensure the integrity of my project and to learn from the leading experts in memory, I conducted an extensive literature review into memory, Artificial Intelligence and existing memorisation solutions for revision.
Insights
Dual encoding of a memory using both visual and written is more effective
Personalising memorisation content to the user can lead to more optimal learning
Nothing in the market which provided Auto generated mnemonics for revision
User Resarch
To learn more about taking professional qualifications whilst also working a full time, I opted to conduct semi-organised interviews with PwC employees to learn about their pain points.
Insights
Memorisation of content was a significant challenge
Wanted a revision tool that was quick to use and required little effort to learn
Users wanted accurate specific content tailored to their qualifications syllabus
(2). Design
Persona Creation
To encapsulate my user research I decided to create a persona in order to be able to refer back to when creating my User Stories and system requirements
Wireframing
Using Figma, I started with low-fidelity wireframes to ensure a solid user experience (UX) for the application. From there, I created high-fidelity wireframes to explore layout options. The design emphasises simplicity and ease of use, featuring a soft colour scheme—light beige backgrounds with dark grey accents—to minimise cognitive overload.
(3). Implementation
Technology Stack
The front end was developed using Flutter, while Firebase Firestore was used as the backend database for storing user data. OpenAI was integrated for generating mnemonics, and GitLab was used for continuous integration.
AI Powered Mnemonics
The app used the OpenAI API to generate personalised mnemonics based on user input, processed through a cloud function to ensure consistent results.
Iteration
By creating MVPs and gathering feedback from stakeholders and users, I was able to streamline the UX by removing an unnecessary step in the mnemonic generation process. Users reported that this optimisation made the process feel faster and more intuitive, making it easier to create mnemonics with minimal effort.
Authentication
Profile Creation - Add User Intrests
Mnemonic Generation
AI Generated Mnemonics
AI Generated Mnemonics
Personalising Content to Syllabus
(4). User Testing
User Study
I conducted a user study with 15 participants to evaluate the effectiveness of my application compared to two leading competitors, Anki and Quizlet. The study assessed how well each application helped users memorise sections of the ACA qualification.
Participants completed an immediate recall test, followed by a second test one week later. Students using Memorise scored 25% higher than those using Quizlet, which ranked second, in the immediate recall test and 35% higher in the test one week later.
System Usability
Using the System Usability Scale, I got the participants of the user study to analyse the applications usability. Memorise scored 87.03 putting it into the excellent category.
Reflections and Acknowledgments
Initially, I developed a different solution for this project: a memory palace creator environment built in Unity. However, after consulting with stakeholders, I decided to pivot and realign my solution with the needs of my target audience. This was a crucial lesson in understanding the importance of designing for users. Following the pivot, I had a much shorter development timeline to learn mobile development and build a functioning application. As a result, I worked long hours every day up to the deadline to complete the project. Though demanding, it was an incredibly rewarding experience.
I would like to thank my supervisor, Professor Pieter Joubert, for his support and guidance throughout this project. He was always available to discuss my work and provide valuable advice.