Making Sense of Data Science: 4 Key Skills Your Data-Driven Company Needs

Posted on September 13, 2017

Frank Kane

Frank Kane

Udemy instructor and Founder of Sundog Software, previously at Amazon and IMDb

data science skills data driven company needs header image

Our increasingly digital world produces a constant stream of data that companies can leverage to gain competitive advantage. By extracting meaning out of all this real-time information, organizations can execute strategic data-driven decisions on just about everything from product development to sales & marketing.

Whether it’s the logs from your website, usage trends about your products, or tracking data from your supply chain – having deep analytical skills in-house can provide insights into your business and help you make the right decisions based on what your “big data” is telling you.

As more companies collect and analyze data, the field of data science remains a hot one. Glassdoor ranked “data scientist” as the top job of 2017, for the second year running. IBM predicts demand for data scientists will grow 28% by 2020.

However, there is a shortage of data science talent as companies struggle to fill these new roles. McKinsey estimated the U.S. gap in data scientists would be around 140,000-190,000 by 2018, resulting in a demand that’s 60% greater than supply.

Corporate learning & development (L&D) can help address this gap by training existing and new data science talent. Although the field of data science is very broad, with many specialties relevant to different situations, here are 4 skills your company should consider focusing on.

1. Data science algorithms

Think of algorithms as a data scientist’s tool chest. There are the many different techniques a data scientist can use to extract meaning from your data. For example, an algorithm called linear regression could be used to predict how much money a customer is likely to spend at your store based on their age. These algorithms are complex and rely on various programming languages. The data fed into these algorithms requires careful preparation and thoughtful design.

For basic data science literacy, entry-level courses can help anyone with some technical and programming or scripting experience to interpret your company’s data. Data scientists often use the easy-to-learn Python programming language, so starting with an online Python crash course such as the Complete Python Bootcamp on Udemy for Business is a great first step. After that, employees can take my course Data Science and Machine Learning – Hands On which explains all of the major data science and machine learning techniques in accessible, plain English while cutting through the field’s intimidating academic notation. In addition, consider the popular Python for Data Science and Machine Learning Bootcamp course on Udemy.

data science skills screenshot

While using Python gives data scientists a lot of power and flexibility, it’s possible to perform data science without Python programming expertise. The simpler R scripting language is an alternative to Python, or an analyst might even use a software package called Tableau that doesn’t require programming at all.

Udemy for Business covers all three options. The popular Data Science A-Z course uses Tableau, and Data Science and Machine Learning Bootcamp with R is also highly rated. The right choice depends on the computer science background of your employees, as well as any pre-existing expertise in Python, R, or Tableau that may exist in your organization.

2. Experimental design

Business leaders who need to interpret data and make decisions based on it should also understand how to design and properly interpret the results of online experiments known as A/B tests. For example, you might use an A/B test to determine the best placement, size, or color of the “Buy button” on an e-commerce website. By displaying different buttons to different customers and measuring their behavior, you can figure out which button will generate the most revenue.

However, it is very easy to “lie with statistics” and make decisions based on correlations that don’t actually reflect causality. Without a strong data science background, you might think one button performs better than the other – but in fact, there wasn’t enough data to know for sure. Or what looks like a definitive result could just be the result of seasonal changes in behavior over the holidays, or the short-term effects of customers seeing something that’s new. A strong education in experimental design can prevent your business from making the wrong decisions, even when they appear to be “data-driven.”

The concepts of T-tests and P-values are critical to interpreting the data that arises from your company’s controlled experiments. They are measurements of the confidence in a result being real and not random. Understanding Bayes’ Rule can also keep you on the lookout for false positives that arise from your data—this is a very real problem in the field of medical screening that leads to unnecessary surgeries and anguish. These are critical skills to have in any data-driven business setting.

My data science course covers these topics, and also how to ensure your incoming data is free of anomalies or fake data. Fake data is just what it sounds like. On a website for example, it’s important to make decisions based on the behavior of real customers, not from search engines, scripts, or hackers trying to steal your data.

3. Machine learning and deep learning

Artificial intelligence (AI) is all the rage lately, but it’s not a panacea. The best way to understand what today’s AI technology can offer your company – or not – is to learn more about it. Deep learning is a subfield of machine learning which focuses on algorithms that mimic learning like a human brain. This is known as artificial neural networks. The concepts of neural networks and “deep learning” are surprisingly simple! They are also surprisingly powerful.

Companies like grocery chains are incorporating machine learning algorithms from driver routing to personalizing online shopping lists or ordering groceries through AI-powered home assistants that remember what you buy regularly and serve up any relevant price discounts. Deep learning is also being used to automatically produce video captions and transcripts, or to translate material from one language to another–which is exciting for any business that produces content for an international audience.

Machine learning is not limited to neural networks; there are more established and effective techniques out there—and most of the data science courses mentioned above will cover them.

Deep learning and AI is, however, its own specialty. The Lazy Programmer offers a suite of well-received courses on this topic at Udemy, and there’s also the popular Deep Learning A-Z: Hands-On Artificial Neural Networks course with great hands-on practice exercises as well. Some specific hot technologies in the deep learning space are open source code libraries like Tensorflow, Theano, and Keras—which will enable you to implement deep learning models with just a few lines of code.

data science skills screenshot

4. Wrangling big data

While it’s great to learn about data science using your own PC, the scale of the data your business generates probably exceeds the computing power that’s on your desktop. To bring these algorithms into production, you need to know how to distribute them across the computers in your data center, or in cloud-based solutions.

The world of analyzing big data is complex, with hundreds of oddly named technologies that work together. My Ultimate Hands-On Hadoop course explains the major components that are out there, and how they all fit together, with lots of hands-on activities. Employees who wish to dive deeper into particular technologies can learn more about the Apache Spark and Elasticsearch platforms with me as well. Both platforms offer scalable solutions for analyzing massive data sets over a cluster of computers.

Data science is changing quickly

The field of data science is rapidly evolving, as cloud computing and Graphics Processing Unit (GPU) computing technologies enable capabilities we only dreamed of just five years ago. Innovations in cheaper and more flexible computing power technologies are enabling AI and machine learning like never before.

However, staying on top of all these innovations can be a challenge. Udemy for Business’ large pool of expert instructors ensures that courses on new technologies are produced as quickly as they emerge. Additionally, the feedback from Udemy’s 17 million+ students means that only the best courses on these hot technologies are surfaced to your organization.

Frank Kane is a Udemy Instructor and Founder of Sundog Software. Frank spent 9 years at Amazon and IMDb, developing and managing the technology that automatically delivers product and movie recommendations to hundreds of millions of customers. Frank holds 17 issued patents in the fields of distributed computing, data mining, and machine learning. In 2012, Frank left to start his own successful company, Sundog Software, which focuses on virtual reality environment technology and teaching others about big data analysis.

More resources

By signing up you agree to our Terms of Use, Health and Wellness Waiver and Privacy Policy.

Thank you!