One of our team members is an avid stock trade and owner of an Amazon Echo Dot. He noticed that if you go on the Amazon Alexa Skills stores, there are no stock market related skills that give you personalized recommendations for stocks or have a rating of 2.0 stars out of 5.0 stars or higher. Thus, Nexa was started.
What it does
Nexa is an Amazon Alexa skill for stock market information and personalized recommendations on your Alexa-enabled devices. Nexa is written in Python and hosted on Amazon Web Services. Nexa primarily uses the NASDAQ API for stock information retrieval. It integrates this data with AWS Lambda in order to provide a simplified, comprehensive user experience. Nexa also utilizes Tensorflow to implement latent factor analysis through SVD models to create personalized stock recommendations based off user portfolios.
You can use Nexa to ask about stock information and personalized recommendations by using phrases such as...
Stock close, intraday high, and intraday low prices
- "Alexa, ask NASDAQ what is the price of Apple?"
- "Alexa, ask NASDAQ for Netflix stock price."
- "Alexa, ask NASDAQ what is the low price of Snapchat?"
- "Alexa, ask NASDAQ what is the high of Nvidia?"
- "Alexa, ask NASDAQ for my portfolio update."
- "Alexa, ask NASDAQ to give me a portfolio update."
- "Alexa, ask NASDAQ for some recommended stocks."
Our two primary challenges when building Nexa were becoming familiar with the Alexa skill development environment and creating an effective model for portfolio recommendations. Because none of the us knew much about developing Alexa skills, it was difficult at first to just get Alexa to respond "hello". Additionally, we also wanted to use very complex algorithms (such as convolutional neural network classifiers) for our portfolio recommendations model, but instead realized the impracticality of such complex models, and opted for a simpler and more effective svd model.
Accomplishments we're proud of
We created a full Alexa skill capable of being used on your Amazon Alexa devices right now. We are also proud of our personalization models and integration with a user's personal stock portfolio.
What's next for Nexa
There were numerous more features that we were planning on integrating into Nexa, but because of time, were not able to. Some of these features include creating a news feature which would update users on the latest new headlines for a specific stock, as well as provide sentiment analysis using natural language processing. Another feature we wanted to implement were real-time calculated support and resistance price levels used by day-traders and swing-trades. We also want to improve our existing features, especially our personalized stock recommendations to possibly use a more advanced classifier. Overall, Nexa strives to combine the power of stock information and personalized recommendations with the power of Alexa and its AI, speech, and voice capabilities.