As many people pivot towards technology careers or target new roles with additional skills, bootcamps are becoming a valuable form of learning in the realm of addition education. I recently completed the Data Science Immersive at General Assembly. I was drawn to the program for its concise and intensive schedule combined with industry professionals guiding the learning path. When working under a tight schedule with such large volumes of new information to consume, its difficult to develop a best practice strategy on the fly.
Here’s my retrospective thoughts on guidelines I would share with anyone looking to begin a data science immersive program:
Data Manipulation.
Its easy to get lost in the volume of algorithms and data concepts that are introduced during a immersive program. The sheer number of versions, variations and turning parameters make it difficult keep up. My expectation was that I would need to become an algorithm guru in order to succeed in a data science role. However, an vital rule to remember is that algorithm are nothing without good data.
Your learning success is limited by your ability to manage and manipulate data. Its easy to get caught up in the excitement of machine learning concepts and neglect the basics. Remember, it all relies on the data. Focus on mastering data types and structures in Pandas and Numpy as your ability to quickly and efficiently arrange and present data is vital building an effective model.
Data Visualization.
A critical step in the data science work flow is exploratory data analysis (EDA). Data visualization helps in understanding the scope and shape of the data set, lending insight into the best approach for addressing the problem.
There are a number of great data visualization packages out there, my favourites include Seaborn in Python and Tableau. Each product has its own work flow and tuning parameters, making the learning greater than you might initially expect. Take the time learn deep on which ever package you choose, the ability to graphically articulate the contents of a data set is a vital skill set for any data professional.
Organization.
General Assembly chose to organise the digital content for the program by week rather than by topic. This made it difficult to refer back to previously learned concepts as we progressed through the program.
If you find yourself in a similar position, I highly recommend labelling and organizing your content by subject matter wherever possible. Much of the skills you learn will be applicable multiple times over, having the ability to quickly reference some previous code or notes will beindispensable as your work through the course.
Notes
In a similar vain to organization, note play an important role in understanding the content you are consuming. From the onset I chose to work in Evernote, a free application that lets you capture notes in one central location. Its biggest strength is the ability to take a variety of forms such as visual, audio and written, across multiple devices. Additionally the package date stamps each note and converts all content to be completely searchable (even images and audio).
Projects
As you are working through the program and learning new skills, you inevitably think of amazing applications for your new found talents. Keep track of every idea that excites and note any interesting problems or projects think you can solve. After completing your program, look to your list for great ideas ways to polish your skills and build out your portfolio.