In this 1.5 hour panel discussion, 5 highly qualified data science practitioners representing either academia or industry share their thoughts on issues surrounding data science, and statistics. The video itself is well worth the 3 hour watch and rewatch.
As I patiently reorganised my frenzied, sometimes multilingual, note-taking, I slowly realised just how many questions, surrounding not only data science and statistics, but also machine learning (ML) and the complex interplay of issues between any of the three, that this video answered for me. Here is the product of that reorganisation, with my own thoughts sprinkled throughout.
Why are so many people talking about data science?
Data science is a nascent field that draws upon many different quantitative disciplines to achieve the goal of extracting meaning from raw, noisy data. It piques the interest of not only businesses, but also consumers and job seekers.
For businesses, the advent of digital age facilitates near-infinite data acquisition. Problem is, they may not even know how to store and process terabytes of information, let alone make sense of all this data. Data science offers a scalable method of extracting valuable business intelligence from the firehose of numbers and words, and eventually a path towards monetisation.
For us the consumers, we may seek to understand the magic behind highly targeted advertising, social media content, and search engine results. How does Amazon deliver such relevant shopping recommendations? How does Facebook feed me such engaging material from my social network? And how does Google give me the result I am looking for, 9 times out of 10? Later, when we uncover the secrets and tricks employed by these companies, we may even wish to police unethical behaviour of data harvesting.
For job seekers, data science is a fluid field short on labour and ripe for carving out niches in different application domains or becoming methodology experts, as Patrick Wolfe explains in the video. Chris Wiggins predicts an further diversification in data science-related jobs down the road: big data companies need data scientists, data engineers, data dev-ops, data product managers, and other data-themed positions. The challenges posed by big data coming in thick and fast demand diverse talents, each responsible for different workflow domains.