1 of 7

418 Teapot

A Marketplace for Open Science Projects

Mustafa HINCAL - Team Lead

Mehmet Akif OZDEMIR - Frontend Developer

Görkem Rıdvan ARIK - Backend Developer

Tunahan KARYAGDI - Backend Developer

Ibrahim UYSAL - AI Developer

2 of 7

Our Purpose

The purpose of our project is to connect projects and individuals who want to contribute

Problems

Manual Management: The tagging process is typically done manually, requiring time and labor.

​

Tags Become Outdated Over Time: Manual tagging can lead to tags becoming obsolete and outdated as time goes by.

​

Inadequate for Emerging Concepts: As new skills, terms, and concepts emerge, older tags may struggle to adapt to these developments.

​

Issue with Project Specifics As projects become more detailed, the tagging system may struggle to accurately categorize them, leading to a loss of functionality.

3 of 7

Solution

And the solution we bring here doesn't completely disregard the past. Leveraging data from past experiences, we have developed a Natural Language Processing (NLP) model.

Here, you are seeing the clustered representation of approximately 15,000 project data points that we have created. In reality, this data is in a dimension that cannot be directly visualized, so we reduced it to two dimensions. We employed Principal Component Analysis (PCA) for this reduction. To cluster the data, we used the KMeans algorithm.

4 of 7

We will match these projects based on user profiles. There are several options for creating a user profile. We create a user profile summary by uploading a CV, and we provide users with suitable projects based on this profile.

5 of 7

This artificial intelligence model, when continuously trained with new data over time, will ensure that it remains up-to-date and capable of recommending the most relevant open-source scientific projects to relevant users. This way, it will maintain its relevance and effectiveness.

6 of 7

Here is the login page. We can login with google and github.

7 of 7

​

    • By augmenting the dataset with more diverse and detailed information, our model can operate even more effectively. This expansion can lead to a substantial improvement in connecting project owners with the most fitting contributors, creating a win-win situation for all stakeholders involved.

​

    • While users can find projects that match their profiles, project owners may have difficulty finding users who are suitable for their projects. After obtaining the necessary data, such a development can be made to address this issue.

​

What can be improved in this project?

Source Code

https://github.com/mustafahincal/marketplace-for-open-science-projects