The Need to predict Stock Prices using both Historical Data as well as Social Media Sentiments and establish a relation between the different methods of predicting stock prices. While doing the Hackathon the long hours of sitting on the laptop without any physical activity showed me that at the various Tech Companies often sit for long hours and thus can cause physical problems in the future.
What it does
The Data Science project is used to predict future stock closing prices using different Machine Learning techniques such as SVM, LTSM, RNN and more. The Fit project leverages the Google FIT API and python on machine in order to lock down the system and give access to the user only when the user has completed a physical task such as walking 'x' steps or so.
How I built it
Using the various ML libraries and framework as theano, keras and tensorflow. Google FIT API, Android Studio have been used to develop the application.
Challenges I ran into
OAuth2 for the Android application and signing for API Calls Selecting the parameters which would actually make a difference in the final Prediction of the Stocks.
Accomplishments that I'm proud of
Ability to approximately predict the market prices on Monday.
What I learned
Markets are not Random and depend on a multiple factors which at times people are very lazy to put together. People are lazy and would start thinking of ways to try and override the setting of the Lock Feature to make the user complete a certain number of steps
What's next for DataScience while Fit
Make it into a complete application, improve the accuracy of the prediction and use more parameters to connect and predict the future prices of multiple stocks. Make application easy to use with the Wearable technology and spread it across to more devices.