How Data Science impacts your life

In their 2012 article “Data Scientist, the Sexiest Job of the 21st Century”, Harvard Business Review credits DJ Patil and Jeff Hammerbacher, early leads of the function at LinkedIn and Facebook respectively, with establishing the modern usage of the term “data science” in 2008. A quick look at Google trends will confirm that its usage has exploded since then.

You don’t have to venture far into the internet to find numerous publications asserting that “data is the digital oil of the new economy” and data science is the tool to unlock its potential, but when we move beyond the hype and the buzzwords, what is left? What does that actually mean? How is data science being applied practically right now in the real world to create value.

What is the true value that data science brings?

More than anything, what data scientists do is make discoveries while swimming in data. -“Data Scientist, the Sexiest Job of the 21st Century”

As I continue to dive deeper into the study of data science, immersing myself in the theory, tools, and techniques one needs as a practititioner, my mind constantly returns to this focusing question of practical application.

So naturally, you can imagine my pleasure at discovering the Super Data Science Podcast, hosted by Australia-based entrepreneur and COO of AI consulting company BlueLife AI, Kirill Eremenko.

Kirill’s motto is “let’s make the complex simple”, and on the podcast he helps to cut through the noise and paint a realistic and easy to understand picture of what an impactful career in data science can look like by interviewing and sharing the stories of data science professionals ranging from machine learning engineers to recruiters, to storytellers to machine learning engineers to visualization specialists and more.

I particularly enjoyed his series of brief 15 minute solo episodes providing a high-level overview of some of the ways data science is being put to use in the industries that impact us each day.

While I certainly recommend listening to the episodes yourself, here are a few of my favorite examples that stood out:

Real Estate

Real Estate is a hugely impactful industry that touches nearly everyone on earth, with global annual revenues in the trillions and cumulative asset values over $200 trillion USD.

Every day, people around the world are asking themselves “How much is my home worth?” or on the other side of the transaction, “How much is the home that I want to buy worth?”

Zillow’s Zestimate combines a ton of data from disparate sources, both structured and unstructured to provide much better appraisal estimates than were available in the past. The accuracy of their model continues to improve over time, currently boasting a median error rate of just 1.9% for on-market homes and 7.5% for off-market homes

The Zestimate? home valuation model is Zillow’s estimate of a home’s market value. The Zestimate incorporates public and user-submitted data, taking into account home facts, location and market conditions.

https://www.zillow.com/zestimate/

Recommendations are another ingenoius use case of data science in the real estate industry. We are very familiar with recommendations from our daily or weekly interactions with online services like Netflix, Amazon, and Google. However in real estate, the amount of data per individual will typically be much smaller, as most people will only perform a handful of these larger transactions in their lifetime.

To deal with this problem, companies like Trulia Insight choose to form groups or cohorts, clustering together the data of similar indivuals allowing them to still make valuable recommendations.

In the episode, Kirill also goes into smart homes, lead identification, and automated property management as part of his data science in real estate summary.

Agriculture

Agriculture is another enormous and impactful industry with worldwide global revenues over 2 trillion dollars, and over 1 billion people employed.

Blue River Technology produces a robot called “See & Spray” leverages computer vision, artificial intelligence, and machine learning algorithms to identify and eliminate individual weeds. They claim this allows farmers using their product to reduce their herbicide expenditure and use by 90% while still removing 80% of unwanted weeds. When you consider that there are now 250 species of herbicide resistant weeds, causing $43 billion of annual loss, you start to realize the potential power of such an innovation.

As hard as it might be to believe, tens of thousands of acres of strawberries and other delicious fruits are still picked by hand every single year. At a time when farmers are facing labor shortages, Harvest Croo Robotics has created a robot that using machine learning and computer vision can produce equivalent daily output to 30 human workers.

Precision agriculture, weather prediction, and biodiversity are some other topics covered in this area.

While these are just a few of my favorite applications mentioned above, other episodes delve into industries as varied as Transportation, Healthcare, Entertainment, Mining, Retail, Construction, Wealth Management, Government, and Education. While they are only touching the tip of the iceberg of what can be accomplished with data science, they really get the imagination going.

How about you? 

Do you have a favorite data science story or application? And if so, what is it?

I'd love to know!