Wow! This sounds really exciting and interesting! To give you a background, I am adept at using R for building machine learning algorithms like random forests, neural networks, xgboost etc. I have also done textual analysis in the past which I presume would be required for reading in data from financial headlines. I didn't quite understand what you mean by a "median term". Would appreciate if you could share more details. If you want a paper, I could do that too using R markdown.
I think this will be a very interesting project to work on and would love to help you out with this. Feel free to discuss this further with me.