Decision Mechanics

Insight. Applied.

How does data science make you faster, smarter, cooler or richer?

By

In a recent interview, Cameron Davies, head of corporate decision sciences for NBCUniversal, discussed the importance of data science being led by the needs of the business.

He discussed “data culture” and the tendency for data scientists to focus on the technology while losing sight of the wider context.

Davies points out that decision-makers want to know

How does working with you, your data, or your new “tool” make me faster, smarter, cooler, or richer? Isn’t that what everybody is looking for?

Data scientists will fail to add value to their businesses until they can confidently answer those questions.

Python in the Economist

By

The Economist has an article devoted to the popularity of the Python programming language. Citibank, Bain & Company and Boston Consulting Group are using it for their data science work.

As Guido Van Rossum, the inventor of Python, reportedly said

Python has become the language of choice for AI researchers […]

Microsoft Research Open Data

By

Microsoft Research have released over 50 free data sets via their Open Data site. They include

  • 38 million tweets from the 2012 US presidential election
  • Profiles of 1 million celebrities (1000 with images)

Are there actually 1 million celebrities now?! Maybe someone can analyze the data to confirm. As in all data science we’ll need to start with firming up our definitions of terms—e.g “celebrity”.

Real-world datasets for learning data science in R

By

R comes with a range datasets that can be used when learning the basics or trying out a new approach/package. mtcars and weather` are popular choices.

However, most of the common datasets are “toy” examples. They are great for practising basic techniques, but are useless when it comes to realistically simulating data science tasks.

The dslabs package provides datasets more suited to exploring data science techniques. It includes

  • admissions Gender bias among graduate school admissions to UC Berkeley (example of Simpson’s paradox)
  • divorce_margarine Divorce rate and margarine consumption data (illustrates spurious causation)
  • gapminder Health and income outcomes
  • heights Self-reported heights
  • murders US gun murders by state for 2010
  • na_example Randomly generated count data with missing values
  • outlier_example Randomly generated adult male heights (in feet) with outliers
  • polls_us_election_2016 Fivethirtyeight 2016 Poll Data
  • research_funding_rates Gender bias in research funding in the Netherlands
  • rfalling_object Simulate falling object data
  • take_poll Models results from taking a poll