Humans are always prejudiced when it comes to investing in stock markets. Investments are made based more on intuition rather than market analysis and historical trends. To solve the woes of the investors, we have developed an application that accurately tracks the performance of the stock of a company. It also finds the correlation
What it does
Stockly provides a one stop solution for the investors who wish to make an exponential gain in the stock market. It accurately tracks the stock performance of the company and sends out signals on real-time basis to buy or sell a particular stock. It also finds the sentiment of the stock in the social media and makes a correlation between the stock prices and user sentiments which further makes the recommendation more accurate.
How we built it
The application can be segregated into two parts. The first part analyses the social network data. Real-time twitter feeds are collected for the desired company and user sentiments are calculated using Natural Language Processing. A positive score indicates a high potential for positive growth and vice versa. The second part of the application deals with the actual historical stock prices. Exponential smoothing curves are generated using Holt-Winters algorithm to predict the stock prices on a running window basis. The last step is making a correlation between the two data pipelines and making a recommendation to the end user that maximizes the profit on investment.
Challenges we ran into
Integrating the multiple technologies into a single application and further deploying it in AWS was the biggest challenge. Initially, we chose R programming as our primary language for analysis, but we had to change the platform completely because of the limitations in scalability and deployment. As a result, we had to re-write the entire back-end in Node.js. Another challenge we ran into was cleaning the data from API's and removing data inconsistencies to make an accurate time-series model of the stock.
Accomplishments that we're proud of
Despite several integration problems that we ran into, we were able to integrate and host the application on AWS. Right from the beginning, we had managed the work and allocated timelines for each task and that helped in incorporating disparate domains such as data science and web design, software development. We ensured smoother communication and file sharing through Slack and it gave us a unique experience in making quick decisions in limited time which is often the case in real-projects.
What we learned
Node.js was one technology which was relatively new to most of us. We also learnt the benefits of having a work-breakdown structure to solve complex problems in smaller chunks. Each of us being from a diverse background got to learn how contribution from one improved the output of another and taking the functionality of the application to a whole new level.
What's next for https://github.com/dfwcrew/stock-analyzerjs
We wish to take this application to the next level by implementing a SMS feature. Whenever the stock price goes above or below the threshold, the application will send SMS to either buy or sell the stock.