Inspiration

As members of a booming society, we have realized that collaboration has become an integral part of any social dynamic. Be it at work, in class or at an event such as this hackathon, meeting new people to collaborate with is a reoccurring theme. We also realized that this process of finding a team can be very tedious, intimidating and inefficient. Therefore, our team came up with a Web Application that eliminates the process of manually finding teammates by using an individuals interests, past experiences and other personal characteristics to match them with others that share similar qualities.

What it does

The user is greeted with a friendly chatbot that surveys the user for basic information about their past experiences, their goals, interests and other such qualities. This data is then stored and analyzed through an AI algorithm (LUIS.ai) that uses sentimental cognitive analysis, intentions and utterances to allot a specific score to each response. These responses are then matched with other profiles stored in the database by ensuring minimal differences between scores.

How we built it

The frontend of the application was build through the React framework for Javascript, including an easy log-in, user-friendly UI and a chatbot. The backend was written in Python and utilized various features within the Microsoft Azure Library. It used LUIS.ai to build a set of intents and utterances corresponding to the answers that are collected from the chatbot. The AI modelled was trained through numerous examples and tests that helped it predict the users' intentions with a high degree of confidence. The entity aspect of LUIS.ai was also implemented to examine the geographical location of the users' educational institute. Using sentimental analysis as well as LUIS.ai, we were able to tally up scores for the most likely intention of the user for each of their responses, which was then used to build a user profile. The user profile is then compared to all the other profiles that are already stored in the database through our own algorithm. Each comparison is done by taking the differences between the scores allotted for each user and optimal compatibility is determined by the lowest difference.

Challenges we ran into

The main challenge I think we ran into is when we tried to implement an SQL Database within which all the users' profiles would be stored. We used Azure Data Studio to manually build a database locally which was then uploaded to the cloud through the Microsoft Azure Platform. Due to the time constraint and lack of previous experience with SQL, it took a lot of our time trying to figure out the best way to implement a database for our Web App. It wasn't until the end that we realized that there are simpler ways to tackle the same issue in a much shorter amount of time, rather than spending all that time an energy to learn a completely new skill. So, essentially most of the challenges stemmed from the fact that we weren't aware of the many effective resources that are available to us which can make hacking a much more convenient experience.

Accomplishments that we're proud of/ What we learned

As a team, we are very proud of the skills that all of us have attained in the past 36 hours while trying to develop such a diverse and effective application. From learning about different Microsoft Azure and AWS frameworks to using AI to implementing SQL databases, each member of our team will be walking away with a strong set of new skills that were attained during this hackathon. We are also very proud of our final product as it is a great, versatile and effective resource that can be used in almost any work dynamic that requires team-building and collaboration; the possibilities are endless!

What's next for Magic Match

Currently, Magic Match is just focussed on building teams for hackathons, however, in the future we would like to see Magic Match being used in various different work dynamics such as offices, schools, sport teams etc. The opportunities for an application like this are endless and we hope to make these opportunities possible in the future.

Share this project:

Updates