Inspiration
There are too many events are happening in the world and it is going to effect stock market positively or negatively, and most cases these events are not unique and outcome on the stock market are not unique too.
What it does
I have created model to predict outcome on market taking into account of event and where it happened and cause/actor of that event.
How we built it
I build this everything on Databricks using python code.
Challenges we ran into
1) I could not get lot historical data, because of compute power. 2) I could not keep trying with live events and live stock market information to see trends live. there is possibility of singing up with NASDQ to sign up for live data and use Databricks DLT to keep consuming live data from GDELT Project for events data NASDQ
Accomplishments that we're proud of
I could merge both data and did some testing on previous data and got about 80% accuracy.
What we learned
1) Model needs more features to come up with better accuracy. 2) History does repeat itself. 3) It is not hard to come with a model on Databricks and tools that are available in open source/market. 4) Market behaves quicker than it used to in the past, so live data is must
What's next
Try to get live data from GDELT events and stock market data and keep tuning the model. Maybe, share it with institutions or NGO's. I figured that this events data could be used beyond stock market. It could be applied in various events like disaster or positive events.
Log in or sign up for Devpost to join the conversation.