We establish communication via email and create a collaborative group chat on WhatsApp. Additionally, we engage with our clients, coordinating and organizing the initial meeting.

We attend our first meeting with IFRC representatives Gefra and Vincent to delve into the project's requirements. Concurrently, we finalize our weekly meeting schedule for future interactions. Post-meeting, we establish a dedicated Teams group and a Google Drive to streamline project collaboration and information sharing.

During this week, we dedicated time to comprehending Human-Computer Interaction (HCI) concepts, initiating our groundwork for the HCI coursework. Delving into our project, we engaged in discussions about user types, laying the foundation for primary design. Progressing further, we successfully crafted two personas and scenarios, tailored to our identified user types. Simultaneously, we conducted various sketches, exploring diverse ideas to enhance our creative process.

To enhance the requirements section of our HCI coursework, we opt for a semi-structured interview approach with our end users. We meticulously prepare a set of questions for the interview and secure appointments with our users. Following this, we refine our sketches and conclude our prototypes, ensuring their readiness for presentation to users and the collection of valuable opinions and feedback.

We systematically conduct and document the user interviews, meticulously capturing their insights. Subsequently, we distill the information provided by users to conclude the requirements section. In parallel, we evaluate our prototypes, integrating user feedback and suggestions, improving the prototypes to align more closely with the users' needs. This iterative process ensures that our final design is well-informed by user perspectives and requirements.

Post the completion of the HCI aspects, we solidify our project requirements utilizing the MoSCoW method and establish a comprehensive timeline for upcoming project milestones. Engaging in technical discussions, we outline the various components of the project, determining the necessary techniques and libraries to employ. Simultaneously, we create a dedicated Github repository to systematically track and monitor our project's progress, ensuring transparency and collaboration among the team members.

This week, our focus shifts to an initial exploration of the selected libraries for our project, aiming to familiarize ourselves with their functionalities. We engage in hands-on experimentation, specifically delving into the scraper, LangChain, and Chroma libraries. This proactive approach allows us to gain proficiency in these tools early on, laying a solid foundation for their effective utilization in the subsequent phases of our project.

Following extensive experimentation with the chosen libraries, we strategically divide tasks among team members and engage in detailed discussions to streamline our workflow. Specifically, Hao is responsible for obtaining texts and corresponding URLs using the scraper, while Lucy takes charge of processing these texts. Leveraging Langchain, we segment the texts into logical pieces, storing both the actual URL and the main website URL. Berken takes charge of uploading this information to the Chroma vector database and subsequently querying it to extract the ten most relevant pieces. The text completion aspect is currently pending due to financial constraints. We plan to initiate discussions with our clients to secure budgets for utilizing OpenAI services.

We successfully complete the programming tasks outlined in last week's plan, ensuring a standardized output for each program component. In alignment with Object-Oriented Programming principles, we meticulously integrate the functions into classes. Subsequently, we amalgamate the three parts to create an initial prototype, which we present to the teaching assistant for evaluation.

In a continuous effort to enhance algorithm performance, we explore various text splitters, aiming for improved efficiency and effectiveness. This iterative process underscores our commitment to refining and optimizing the algorithm for optimal results.

Commencing preparations for the pitch, we collaboratively craft a compelling PowerPoint and script to effectively convey our project. Simultaneously, we initiate the development of our project website, the blog, hosting it on the university server. While the website is in its early stages, we have successfully uploaded preliminary content, laying the groundwork for a comprehensive and informative online presence.

Following thorough preparation, we are ready to record our presentation. Upon the initial attempt, we collaboratively adjust the script to ensure it aligns with the 2-minute timeframe. After submitting the pre-recorded video, we strategically divide tasks for the live presentation and engage in collective rehearsals. Simultaneously, with regard to the blog, we complete and upload essential content covering requirements, HCI, design, and research sections to the website hosted on the university server.

This week, we engage in discussions with our client regarding the funding for utilizing the OpenAI API for the text completion aspect of the project. Securing the secret key for the OpenAI account with a $15 usage limit, we initiate the coding process for the text completion component. Simultaneously, we commence the implementation of the project's front end, adhering to the design principles outlined in our prototype. In a collaborative decision, we opt to utilize the React framework for the front-end development. Additionally, we organize and refine our past development logs to maintain a comprehensive record of our progress.

We solicit sample questions from our clients, focusing on topics such as climate change, to assess the program's effectiveness. To ensure a comprehensive answer, more approved sources related to those questions are incorporated. Concurrently, we engage in discussions concerning enhancements to various aspects of the web interface.

Simultaneously, we leverage the OpenAI API to enhance the precision of text segmentation. Recognizing the trade-off between accuracy and speed, we implement threading mechanisms to address any potential latency issues. Additionally, we contemplate the integration of a simple login system to bolster privacy measures and maintain a more detailed record of historical questions.

In terms of project management, we meticulously organize our GitHub repository, consolidating redundant branches for a streamlined workflow. To enhance accessibility and user experience, we revamp our documentation website by introducing separate pages and incorporating detailed weekly logs.

Over these two weeks, our primary focus is on enhancing the front-end design aspect of our project. Following client feedback, it has been determined that the inclusion of a landing page is essential. Additionally, clients have suggested incorporating functionality that allows users to select specific answers during an interaction, leading to the generation of a PDF report.

To accommodate these requirements, we are designing distinct pages within the front-end interface. This approach ensures clarity in user navigation and streamlines the overall user experience, ultimately aligning with the project's objectives and client expectations.

For these two weeks, our primary emphasis has remained on frontend design. Following discussions with our clients, we've solidified the components that the final prototype should encompass. These components comprise a landing page, a login page, the main query interface with history, a data management page, and a FAQ section.

In comparison to our initial design, the expanded content scope renders the project more comprehensive and user-friendly. With the design confirmed, we've opted to utilize the React framework to develop each web page independently. This approach ensures that each component is built to the highest standard and can seamlessly integrate into the cohesive final product.

This week, we implement several frontend enhancements based on client feedback. Additionally, we've made our website multilingual, now supporting English, French and Spanish, to ensure accessibility for users of diverse language backgrounds. Following this, we seamlessly integrate the backend with the frontend, enabling direct usage of the conversation page for querying and obtaining answers. Users can now also access historical questions. However, we've encountered a few bugs post-integration and are diligently working to resolve them.

During the concluding fortnight, we seamlessly integrate various web pages, culminating in a comprehensive and fully functional website, subsequently deploying it via Netlify. Addressing encountered bugs becomes paramount to ensure the website's usability. Additionally, we craft a presentation complete with demonstrations to elucidate our system's functionality to a broader audience, including those less versed in technical jargon.

Simultaneously, we enhance our documentation by finalizing both the documentation website and development logs. With each member completing their individual reports, we cap off this project successfully, delivering a valuable and accessible socio-behavioral tool.

That concludes our journey with this project! On retrospect, through the highs and lows of the project, it was still a valuable learning experience. We've learned to work collaboratively, communicate effectively, and manage our time efficiently. We've also gained valuable insights into the development process, from ideation to implementation.

To all those that supported us, we would just like to say a big thank you! To our client, IFRC, for providing us with this opportunity and guiding us throughout the project. To our teaching assistants, for their guidance and feedback. And to our team members, for their hard work and dedication.

- And that's Team 35 2023-2024 signing off!