My Data Science Bootcamp
The first of December last year, I was raising awareness for World AIDS Day. This year I used my mobile workstation’s GPU to train an artificial neural network six times faster than by CPU processing alone. Both were personal challenges with the power to help people, and both are memorable for the way they pushed me outside my comfort zone.
This week I will present my final Capstone project and graduate from the Data Science Immersive program I began last fall. My virtual cohort of only 12 is dispersed across the western United States and, while there has been little time to chat about our personal lives, it is clear that these career transitioning individuals are some of the brightest, hardest working, and most dedicated professionals I have ever had the pleasure of collaborating with directly.
General Assembly runs a tight ship. The program is rigorous, covering a diverse assortment of selected topics in the field of Data Science ranging from command line and shell scripting to web scraping and API request handling. The active hours currently displayed by my Windows machine are 8:00 AM to 2:00 AM Mon-Thurs, and they’re not much better on the weekends. Some might say I haven’t been getting enough exercise, or sleep, but I found a flow early enough in the bootcamp that I could adapt to the intensity as quickly as it unfolded, and for that I’m grateful. In particular, taking a deep dive into the world of Machine Learning granted me the opportunity to rediscover my appetite for data-driven decision making, problem solving, and quantitative reasoning. I have found that data projects transcend industries, and offer a balance of powerful business intelligence(BI) solutions with an avenue for driving meaningful change in the lives of real people.
My data portfolio has grown to include a few projects I am quite proud of, as well as some I am excited about taking deeper. Specifically, I look forward to imparting newfound imputation techniques with Scikit-learn and optimizing my position towards the top of Kaggle leaderboards. Additionally, a principal component analysis(PCA) will serve well in the regression of the sale price of homes in Ames, Iowa, where a number of predictive real-estate features can easily introduce high variance. Tableau, an intuitive BI solution for visualizing data efficiently, has a free public version that will be put to good use as well. Creating an interactive visualization, representing the complex relationship uncovered between 12th Grader SAT performance and California School District equitability, is a communication essential. Furthermore, the classification of hundreds of thousands of user submissions, originating from one of two similarly themed subreddit discussion groups, suffers from a frequentist’s approach. Words are not independent within a given sentiment and additional data regarding the context in which they occur is likely to improve the model. Bayesian inference, in combination with unsupervised learning techniques such as clustering algorithms, have the potential to illuminate subtle relationships in Natural Language Processing(NLP) yet unseen. Finally, deploying my dog breed classifier as a web app, with Heroku, where users can upload images of their own dog might be the final touch needed to move on to the next phase of my professional journey, the job hunt.
As much as I enjoyed the technical challenge of deep project work, my excitement for Data Science goes beyond the science and technology. I care a great deal about my local community and feel a calling to make the world a safer place. As I launch my job search, I hope to work with local businesses actively benefiting underserved communities by developing advanced business strategies with insightful forecasts. With the curfew orders currently imposed surrounding the most recent surge in positive COVID-19 test cases, many small businesses are fighting to stave off permanent closure. This is an unusual time in the world, full of uncertainty, and by placing information in the hands of business leaders I hope to use my expanding bag of data science, analytics, and engineering tricks for good. The time has come for me to graduate from aspiring Data Scientist to Jr. Data Scientist, working with real data, and keeping real problems where they belong, in the forefront of my mind.
