NioC | Devpost

Inspiration

This project was inspired by our very own interest in cryptocurrency and stocks. With the current pandemic as a global challenge, technology has become more and more integrated into our everyday lives. This has sparked a level of curiosity for many individuals to push into further digital activities such as cryptocurrency. While a lot of people have gained interest in this field, it is often referred to as a "mystery" and a difficult venture to take on. Our web application, however, is here to solve that problem. NioC helps beginners gain insightful understanding and knowledge of trends and predictions related to stocks and cryptocurrency.

What it does

NioC is a conservation AI and tool for data analytics for cryptocurrency and teaching those who may want to learn more about the technology. as a result, giving users inform

How we built it

NioC uses historical cryptocurrency data to gather with Facebook's time series forecasting model prophet to forecast the future prices of cryptocurrency in the future and it also computes the historical volatility of a cryptocurrency.

Challenges

-Initially, we had planned to use the google cloud platform for our chatbot, however, due to complication we had to switch to an open-source alternative, Rassa. Rasa proved to be much more difficult to implement than we had planned, this prevented us from integrating the chatbot into the web app.

-To perform the forecasting of cryptocurrency prices. We had decided to use Facebook's Prophet algorithm. This was much more difficult than expected as installing Prophet on windows was almost impossible to do.

-We also planned to use perform time series classification on the data. There were many issues while attempting to do this. The major issue was pre-processing the data. We needed to separate the data into windows of an arbitrarily chosen time, perform a rolling average on the windows, and then label the data based on its neighbors.

Accomplishments that we're proud of

We took an ambitious task and work relentlessly throughout this hackathon. We are proud of the hard work we each contributed and how well we worked as a team. We are proud to present a variety of data visualizations representing various cryptocurrency and their trends, as well as being able to ideate such a powerful project.

What we learned

We learned that in heavy time constraints, you have to make some sacrifices. Because of unexpected setbacks, the conversational Ai scope had to be lessened considerably.
We learned a lot more about cryptocurrency and its place in the future.
We also learned about conversational Ai and how tedious it is for just an initial model. Despite this, we had a lot of fun doing our section. When we think about it, the effort is worth it because once the initial model is completed, it can learn from having conversations with users.

-learning about prediction in time series was amazing. It is so cool how this can be used in different ways to even predict the rise and fall of currency prices.

What's next for Untitled

The next step is to complete the Conversational bot to be able to teach people about cryptocurrency and so that is able to interpret the graphs for customers. We would also like to upgrade the bot to be able to autonomously train other bots to buy and sell cryptocurrency on a live market. Furthermore, we would like to perform a more in-depth analysis of the cryptocurrency data by combining other types of data from social media platforms, etc

Google Drive to Video https://drive.google.com/drive/folders/1TWjkf6SmcNagZhT8cLqtjXraAEfVowxs