Written by Brandon Sandhu on 13/07/2024
The objective of this initiative was to gain insight into the professional experience of a quantitative researcher. I was assigned four tasks, which entailed comprehending the financial terminology relevant to each task, transforming the problem into a data science context, devising a solution, and subsequently addressing the original problem with the formulated solution. For more information, I refer the reader to Job Simulation Website.
I provide files to the original solutions I made whilst working on each of the problems.
In task one, I gained an overview of commodity storage contracts and learned how to extrapolate data from external feeds to provide granular insights. Additionally, I wrote code that analyzed data, taking a date as input and returning a price for both past and future estimates.
Task One Jupyter Notebook. For this task the following dataset was used: Nat_Gas.csv
In task two, I learned how to write a function that takes particular inputs and gives back the value of a contract. Additionally, I created a prototype pricing model that could undergo further validation and testing before being put into production.
Task Two Jupyter Notebook. For this task the following dataset was used: Nat_Gas.csv
In task three, I learned how to choose appropriate independent variables from a data set to accurately predict the outcome of a chosen dependent variable and understood the importance of using available data to predict customer trends and actions. Additionally, I built a model using Python to estimate the probability of default for a borrower.
Task Three Jupyter Notebook. For this task the following dataset was used: Loan_Data.csv
Finally, in task four I learned how to apply statistical formulas to business solutions and understood the importance of breaking down a large dataset using machine learning methods. Additionally, I deployed detailed Python code to strategically bucket customers with various FICO scores to narrow in on the probability of default.
Task Four Jupyter Notebook. For this task the following dataset was used: Loan_Data.csv