Sean Browne

About Me

For 15 years I have worked in Oil and Gas, mainly Oilfield Services focused on Directional Drilling. In that time I have done a little bit of everything. I entered the industry in an entry level finance role, balancing field tickets prior to billing them. I moved into operations supporting directional drillers in the field as a well design engineer, then later by managing the competency system as a Development Coordinator. I spent time traveling around the United States as a technical training specialist teaching new hires the in’s and out’s of Measurement While Drilling Systems and Well Design. Ultimately, I leveraged all of that into becoming an Account Executive working in the business development group. All while dodging, dipping, ducking, and all the other action verbs associated with demonstrating enough value to make it through, multiple downturns in the industry. Then I found myself caught between a price war in the Eastern Hemisphere and a Global Pandemic.

After the dust settled, I looked around for my next path forward. I reflected on my MBA program and what within it was exciting to me. My data analysis studies were the first thing that came to mind. Data driven decision making was just starting to gain traction within my business unit as I was completing that portion of my studies, so it was perfect timing then. The outlook and recession-proof nature of the work seemed to make it perfect timing for a pivot now. Diving into this subject also presented an opportunity to also look at the front-end development aspect, which has also become an interest. So here I am, building a portfolio to showcase what I have learned and what I am capable of doing for the company that ultimately hires me.

Projects

Music Recommendations

Utilized the 1 Million Song dataset to create a music recommendation program based on overall popularity(naive/unpersonalized) and user selections(personalized/collaborative filter). Recommendations can also be made by song title using the same logic as the user selections. Correlations were made using Pearson Correlation Model.

Recommendations and the ability to make them are a cornerstone of most, if not all, media services regardless of medium (books, music, movies, etc.). I was interested in “peeking behind the curtain” to get an idea of how this worked. The next step would be to apply the program to the full dataset, develop a front-end for users to request recommendations based on songs, playlists, or artists, and finding more data to incorporate into this dataset to strengthen recommendations.

Techniques Used and Dependencies

Collaborative Filtering
Machine Learning
Python
Jupyter
pandas
numpy
scikit-learn

Fandango Rating Study

Exploration of whether changes were made to Fandango’s movie rating system after 2015 article stating it might be inflating ratings. This project was a means of testing what I had learned after working through fundamental statistics.