2017 Business Science Blog In Review

The Business Science Blog has grown tremendously over the past year, fueled by data scientists hungry to learn how to apply data science in business applications. It's been really cool to watch and interact with everyone, whether its been feedback via our Disqus comments or emails through our contact form. It's led us to develop Business Science University, a new online education platform that will offer Virtual Workshops, which is actively being built as we speak. It's also helped us grow from all of the feedback we've received. To those that may be new to the blog, here's a short year-in-review of the most popular blog posts of 2017, which is meant to be a curated guide to those that are interested in the topics we wrote about this past year. Enjoy!

Learning Data Science

Six Reasons To Learn R For Business - A great post for beginner data scientists or BI experts that haven't seen what the R statistical programming language has to offer. We are following it up shortly with an article on how to learn R efficiently and effectively.

Advanced ML How-To's

HR Analytics: Using Machine Learning To Predict Employee Turnover - Our most popular article and the inspiration for our first BSU Virtual Workshop. Exposes the reader to automated machine learning with H2O and explaining black-box models with LIME on a real-world application of predicting and explaining employee turnover.

Customer Analytics: Using Deep Learning To Predict Customer Churn - We used the new keras package in R to build a deep learning model that predicts customer churn. We also used a new preprocessing package, recipes, that makes the numerous data preparation steps a cinch.

Sales Analytics: How To Use Machine Learning To Predict And Optimize Product Backorders - This article uses H2O and LIME, but extends the analysis to optimize Expected Value, a concept that is critical for business.

Tidy Time Series Analysis

We did a 4-part series on time series analysis using tidyquant, our most popular package. We showed off a number of powerful tools such as rolling functions, rolling correlations, and lags/autocorrelations.

• Tidy Time Series Analysis, Part 1 - Using tq_transmute() to apply functions to time intervals
• Tidy Time Series Analysis, Part 2: Rolling Functions
• Tidy Time Series Analysis, Part 3: Rolling Correlations
• Tidy Time Series Analysis, Part 4: Lags and Autocorrelation

Demo Week

Demo Week was a cool way to expose new users to some of our time series packages including tidyquant, tibbletime, timetk and sweep. It was a whirlwind with 5-packages in 5-days.

• Day 1: tidyquant for getting and manipulating financial and economic data
• Day 2: Time Series Machine Learning with timetk
• Day 3: Tidy Forecasting With sweep
• Day 4: Tidy Time Series Analysis With tibbletime
• Day 5: Time Series Machine Learning with h2o and timetk

Wrap Up

We had many more blog posts and happenings during 2017, but unfortunately, they couldn't all make it into this list. If you're interested in applying data science in business applications, please subscribe to the Business Science Blog. We have a lot planned for 2018 including Business Science University, a revolutionary new platform for teaching data science in business. Enrollment is open now, which will keep you updated and gain you access to early-bird pricing. We are really excited to start the new year with a bang!

About Matt

Matt's the founder of Business Science, a consulting firm specializing in applying data science to business. Matt regularly contributes articles to the Business Science blog, a great resource for those looking to learn. He's also working on Business Science University, an online educational platform designed to teach novice-to-intermediate data scientists how to implement advanced machine learning algorithms and build ML-powered web applications within their organization. The first Virtual Workshop will be available in early 2018. You can learn more about Matt and Business Science by connecting on social media:

• LinkedIn: Matt // Business Science
• Twitter: Matt (@mdancho84) // Business Science (@bizscienc)

Never stop learning.