Inspiration
We came in the hackathon knowing we should choose a topic that isn't too easy to create. We tried challenging ourselves to create something cool but useful at the same time. When we heard that Goldman Sachs is sponsoring the event and that we will be able to use their API we knew we wanted to do something finance related using that API. Thats why we came up with the idea to use Neural Networks to predict the upcoming stock prices to suggest if a user should invest in a business or not. We also liked the idea of using Alexa to be the abstract layer between the user and the functions of the system.
What it does
The system uses complicated Neural Networks to predict the future rises and declines in the stock prices. The user asks alexa if he/she should invest in a specific company and then alexa will take the name of that company and see if the answer for that specific company is worth investing in.
How we built it
We started by taking all the data from the GS(Goldman Sachs) API. Then we made the neural net with the help of various different sources. We made up some data sets to test the neural net and then we linked it with the GS API. Finally when the neural net predicts if the investment is worth it or not, we upload it to the firebase and then alexa reads that and outputs if its worth it or not.
Challenges we ran into
Time restriction was a big challenge since it didnt allow for fine tuning for the neural network and small details in general.
Accomplishments that we're proud of
The neural network is impressive for us since we never done anything before that is similar to it.
What we learned
A lot about Neural Networks and how data is processed.
What's next for Marsupial
More tuning and better predictions. As the data increases the system gets better.
Log in or sign up for Devpost to join the conversation.