CollabMarket

Home page
Problems Market Place
Chat function with auto translation

Inspiration

MSMEs in the APEC region struggle to expand their reach into the international marketplace due to the lack of access to up-to-date information, resource constraints and communication barriers. Therefore, we have developed CollabMarket, a comprehensive marketplace and collaboration platform to facilitate seamless communication, resource and knowledge sharing and collaboration among MSMEs in the APEC region. By breaking down international barriers and providing MSMEs access to vital resources, we empower them to unlock their full economic potential in the global marketplace.

What it does

Our key features include:

Marketplace: Businesses posts a request and MSMEs bid to fulfill the request
Collaboration platform: MSME looking to collaborate with another enterprise proposes to collaborate via a call for collaboration
Chat function + auto translation: Users have the option to view messages in original form or their preferred language. With this function, we bridge the communication gap between MSMEs from different cities and countries, which allows them to seamlessly communicate with another MSME or business from another country, establish valuable connections globally

How we built it

Backend

We use FastAPI, a modern, high-performance web framework with Python to build the API service, and data is stored in a PostgreSQL database. We use Docker and Docker Compose to containerize our backend, which provides a standardized and reproducible environment, ensuring consistency across development and deployment phases, streamline the integration of different services, and offer scalability options.

For the chat function, websocket is utilized to provide a real-time and bidirectional communication channel between the client (web browser) and the server.

Translation: Instead of relying on an external API for translation, we run machine learning models on our backend to translate every incoming message to the preferred language of the receiver. The translation system consists of 2 components:

fasttext model for models for language identification, which can recognize 176 languages.
M2M100 pretrained model and tokenizer - a multilingual encoder-decoder (seq-to-seq) model trained for Many-to-Many multilingual translation, which can directly translate between the 9,900 directions of 100 languages.

=> Whenever a message comes in, it will be passed into the fasttext model to classify the language that message is written in. It will be used to set the source target of the M2M100 tokenizer, which tokenizes the message, then sends the tokens into the M2M100 translation model.

Frontend:

Our project's frontend was developed using React, a powerful JavaScript library for building user interfaces. By utilizing React, we created reusable UI components, resulting in a highly efficient and modular development process. To ensure a consistent design, we employed CSS and CSS-in-JS libraries like styled-components and CSS modules, allowing us to encapsulate styles within components. To guarantee the quality of our frontend, we conducted thorough testing. Our rigorous testing approach encompassed functional testing, responsiveness assessment, and cross-browser compatibility checks. This ensured that our frontend components performed seamlessly across different browsers and devices.

Challenges we ran into and limitation of the project

Using Docker and Docker compose, we deployed our backend on a remote server quite easily. However, due to the time and scope of the project, we couldn't get an SSL/TLS certificate for the domain. Therefore, we encountered a lot of problem deploying our frontend since most of the hosting service only accept https connections. After a lot of experiment, we managed to connect backend and frontend in production using a http middleware. However, we could not get the websocket working as expected. The real-time messaging function works well with backend deployed in production and frontend in development environment, but doesn't work well when frontend is deployed on the cloud. We could have tackle this problem if we have more time and resources.

What's next for CollabMarket

Explain how we have built fully functional backend but haven't integrated frontend Much of our backend functionality has been implemented and is even publicly available to anyone interested. These functionalities include: basic authentication, sending requests to create, get, update, and delete problems, collaborations, problem bids, and collaboration requests, 2 way real-time communication for the chat function, and publically available translation service. However, limited by the lack of expertise and time constraints we have not yet integrated these functionalities to our current version of CollabMarket. Moving forward, we hope to fully implement these features on the frontend too. End-to-end business pipeline, from initiation of project to the delivery of solution

As the saying goes, you can bring the horse to the water but you can’t force it to drink. But you can certainly make it easier for it to reach the water. With that in mind we hope to continue building features that will further enable MSMEs to participate in international markets. These features will encompass lightweight versions of tools essential to the business pipeline; from initiation of project to delivery.

A key feature that we would like to implement is payment integration for our problems marketplace. We strongly believe that such a feature will formalize the entire solution and help facilitate trade in a seamless and efficient manner.

Conclusion

CollabMarket embraces the key theme of APEC App Challenge:

Resiliency:

When MSMEs are able to collaborate with one another, share resources and reduce the costs due to economies of scale, this could potentially increase their financial resilience

Sustainability:

Resource efficiency: Fostering collaboration between MSMEs enables the sharing of resources and expertise. This can potentially lead to better utilization of existing resources, reducing waste and improving overall resource efficiency.

Social and economic development: As the platform provides opportunities for MSMEs to bid on problems to solve, the marketplace can potentially enhance the competitiveness and growth of these enterprises. This is turn can contribute to job creation, income generation and poverty reduction within the local community

Inclusivity:

Through our platform, micro and small businesses can collaborate to share resources and also increase their production and benefit from economies of scale. Supports whole spectrum of MSMEs from micro-enterprises to mid-sized firms