Inspiration
We have always been excited by the Stock Market and its unpredictability. This project is an extension of our curiosity as to how the changeability rolls into play.
What it does
InSight takes in a dataset and plots it against a timeline - mapping important incidents and inflections in the graph along the way using a very reliable algorithm. The algorithm then matches the important dates with plausible articles and resources that might explain how the change came about. InSight gives users unparalleled insights into data and resources to corroborate their findings.
How we built it
We used NASDAQ's Real-time stock market API and The New York Times Article Search API to build the backbone of the current version of InSight. We analyze the entire graph using time derivatives of the absolute values of the graphs over varied periods and mapping them to dates that are more likely to be important in the history of a particular stock. These dates are then matched to The New York Times Articles around the same period which mention the stock in them. Using the API, we are able to return the name of all such articles and also their links to the user for easier access. Thus, InSight makes it very easy for the users to research about historical stock data.
Challenges we ran into
The limitations that The New York Times imposes on the use of their API: We can only monitor one call per every ~0.8 seconds which makes the algorithm slow.
The lack of helpful documentation for the same also slowed us down.
Moreover, there were certain limitations on how we can interact with Jupyter Notebooks that made it harder to achieve the visualizations we made.
Accomplishments that we're proud of
Having very little experience programming apps and integrating APIs from the ground up, we had to teach ourselves to do everything. I think we are extremely proud of the fact that we knew when to seek out help and that we never let the project fall apart over whims. We are really happy about implementing the exact plan that we set out with even though it seemed extremely daunting, to begin with.
What we learned
We most certainly learned that mindset guides progress and that there is no shame in being a beginner or seeking out help. On the contrary to our previous believes, seeking out help from amazing mentors and sponsors has been by far the most enriching experience in the event.
What's next for InSight
We definitely want to integrate more dependable sources into the app and support more datasets so that InSight is able to help more people out. This is the age of Machine Learning, therefore, we would like to implement machine learning algorithms to give users better insights into the data and also to get better predictive points to pull out more information from.
Log in or sign up for Devpost to join the conversation.