Big Data Management & Analytics
In this course you will learn the fundamentals of Data Science and how to approach Big Data
management problems. Specifically, you will be able to:
• Describe the main steps of the Cross-Industry Standard Process for Data Mining
(CRISP-DM);
• Distinguish among the four different Data Science methods covered in class;
• Given a specific Big Data problem, apply the adequate Data Science method to solve it,
and evaluate its performance (both practical and statistical);
• Using the frameworks discussed in class, recognize the ethical dilemmas in collecting
and analyzing Big Data;
• Evaluate the ethical position of a firm in a specific data collection situation;
• Apply the Big Data tools architectural principles to formulate a solution for real-
world scenarios.
Session 1. Session 5.
Introduction & Data-Analytics Thinking Additional Data Science Topics
Session 2. Session 6.
Data Collection and Ethics Big Data Architecture, Engineering & Tools
Session 7.
Session 3. The Future of Big Data
Introduction to predictive modelling
Session 4.
Model Fit, Overfit and Evaluation Written Test
--- Interloop – Optional Clarification
Introduction & Data-Analytics Thinking
Session 1 – Part I
- “Doing Big Data Without Knowing It” – Maybe size doesn’t matter?
- The Goldmine anecdote: – “We’re sitting on a goldmine of data that we don’t know what to
do with”
- The ‘Meh’ anecdote: – “Will the price go up or will the price go down?” à Business
understanding/problem first
The future is already here, it’s just unevenly distributed” – William Gibson
1. Tailoring to customer preferences
,– Facebook Newsfeed, Amazon product recommendations, Google search results, Political
advertisements
2. Running massive experiments:
– Cool blue’s product page, Netflix recommendations, Voting during elections
3. Pattern recognition (deep learning!)
- Shazam
- Google job candidates
- Automated photo tagging
- Driving patterns in cities
4. Text mining
- Customer opinions of products and companies
- Input for developing new products
- Find untapped market segments by mining customer reviews
- Online buzz to forecast new product sales
5. Combining data that hasn’t been combined before
- Using Twitter to map global (un)happiness
- Predicting flu based on Google searches
- Detecting anthrax attacks using grocery sales
- Using GPS data to predict (and contain) Ebola outbreaks
- TomTom can predict air pollution
- RFID to predict teamwork effectiveness in R&D productivity and surgical procedures
Common Elements in all of these:
The V’s of Big Data
1. Volume; Best understood
2. Velocity; input data is reasonable understood, but velocity of output
(analysis/visualization/decisions) much less understood
3. Variety; coolest new applications
4. Veracity; elephant in the room that no one acknowledges
Future that will be here (soon)
• Better data and better analytics should lead to better decisions...
, Not happening overnight, but trend is clearly there – Data collection tools and data-analytic
tools need to co-evolve but this takes time!
• Example: evolution of data and analytic concepts in the NBA
– Drawn from BIM MSc. thesis Maikel Ooms (2016) NBA example
Previous data was still data in the horizontal plane: x-y coordinates.
But the z-coordinate (height) obviously plays an important role in basketball as well... à 3D
Data
- Natural science and medical science have the telescope, microscope, thermometer,
MRI-scanner, ECG, etc
à With Big Data and Business Analytics, business now has all of these ROLLED INTO ONE!
“Brave New World”: Utopia or Orwellian/Huxleyan/Eggerian Dystopia?
- Ethics and (algorithmic) accountability essential elements
IMPORTANT LESSON 1: BIG DATA AND ANALYTICS BY ITSELF DON’T GIVE
BUSINESS VALUE à ALWAYS PUT THE BUSINESS DECISION AT THE CENTER
Analytics in Business
Different typed of Analytics
Different types of analytics have different rules for humans:
The previous wave of analytics was primarily data-driven: New big data
We can measure things we could not measure before
Analytics 1.0 Traditional Analytics
Analytics 2.0 Big Data
Analytics 3.0 Fast Business Impact for the Data Economy
, Digging in the Goldmine…
Just because you can measure it does not make it useful
à What decision will you make based on this (new) data?
Analytics 3.0 will have the decision at the center of it all!
- A seamless blend of traditional analytics and big data
- Analytics integral to running the business; strategic asset
- Rapid an agile insight delivery
- Analytical tools available at point of decision
- Cultural evolution embeds analytics into decision and operational processes
MANAGING BIG DATA AND BUSINESS ANALYTICS IS DIFFERENT FROM REGULAR IT
INITIATIVES
1. Big Data has several unique challenges, as well as new opportunities
– Four V’s
2. Big Data and Analytics projects cannot be managed like regular software projects
– Much more like R&D than building an application
3. Analytics (or data science) is a very different way of thinking about data than normal statistics
– Decision-Oriented instead of Model-Oriented
4. Data-analytic thinking can become so integrated in a company (and in several leading companies it
already is!) that it transforms the company
– Analytic-driven companies do better!
Terminology
• Data strategy refers to the strategy for creating and capturing value from data (either existing
or new)
– Requires Data-Analytic Thinking: being able to assess whether and how data
can improve performance
• Data science involves principles, processes, and techniques for understanding phenomena
via the (automated) analysis of data.
• Data mining is the extraction of knowledge from data, via technologies that incorporate
the principles from data science
In this course you will learn the fundamentals of Data Science and how to approach Big Data
management problems. Specifically, you will be able to:
• Describe the main steps of the Cross-Industry Standard Process for Data Mining
(CRISP-DM);
• Distinguish among the four different Data Science methods covered in class;
• Given a specific Big Data problem, apply the adequate Data Science method to solve it,
and evaluate its performance (both practical and statistical);
• Using the frameworks discussed in class, recognize the ethical dilemmas in collecting
and analyzing Big Data;
• Evaluate the ethical position of a firm in a specific data collection situation;
• Apply the Big Data tools architectural principles to formulate a solution for real-
world scenarios.
Session 1. Session 5.
Introduction & Data-Analytics Thinking Additional Data Science Topics
Session 2. Session 6.
Data Collection and Ethics Big Data Architecture, Engineering & Tools
Session 7.
Session 3. The Future of Big Data
Introduction to predictive modelling
Session 4.
Model Fit, Overfit and Evaluation Written Test
--- Interloop – Optional Clarification
Introduction & Data-Analytics Thinking
Session 1 – Part I
- “Doing Big Data Without Knowing It” – Maybe size doesn’t matter?
- The Goldmine anecdote: – “We’re sitting on a goldmine of data that we don’t know what to
do with”
- The ‘Meh’ anecdote: – “Will the price go up or will the price go down?” à Business
understanding/problem first
The future is already here, it’s just unevenly distributed” – William Gibson
1. Tailoring to customer preferences
,– Facebook Newsfeed, Amazon product recommendations, Google search results, Political
advertisements
2. Running massive experiments:
– Cool blue’s product page, Netflix recommendations, Voting during elections
3. Pattern recognition (deep learning!)
- Shazam
- Google job candidates
- Automated photo tagging
- Driving patterns in cities
4. Text mining
- Customer opinions of products and companies
- Input for developing new products
- Find untapped market segments by mining customer reviews
- Online buzz to forecast new product sales
5. Combining data that hasn’t been combined before
- Using Twitter to map global (un)happiness
- Predicting flu based on Google searches
- Detecting anthrax attacks using grocery sales
- Using GPS data to predict (and contain) Ebola outbreaks
- TomTom can predict air pollution
- RFID to predict teamwork effectiveness in R&D productivity and surgical procedures
Common Elements in all of these:
The V’s of Big Data
1. Volume; Best understood
2. Velocity; input data is reasonable understood, but velocity of output
(analysis/visualization/decisions) much less understood
3. Variety; coolest new applications
4. Veracity; elephant in the room that no one acknowledges
Future that will be here (soon)
• Better data and better analytics should lead to better decisions...
, Not happening overnight, but trend is clearly there – Data collection tools and data-analytic
tools need to co-evolve but this takes time!
• Example: evolution of data and analytic concepts in the NBA
– Drawn from BIM MSc. thesis Maikel Ooms (2016) NBA example
Previous data was still data in the horizontal plane: x-y coordinates.
But the z-coordinate (height) obviously plays an important role in basketball as well... à 3D
Data
- Natural science and medical science have the telescope, microscope, thermometer,
MRI-scanner, ECG, etc
à With Big Data and Business Analytics, business now has all of these ROLLED INTO ONE!
“Brave New World”: Utopia or Orwellian/Huxleyan/Eggerian Dystopia?
- Ethics and (algorithmic) accountability essential elements
IMPORTANT LESSON 1: BIG DATA AND ANALYTICS BY ITSELF DON’T GIVE
BUSINESS VALUE à ALWAYS PUT THE BUSINESS DECISION AT THE CENTER
Analytics in Business
Different typed of Analytics
Different types of analytics have different rules for humans:
The previous wave of analytics was primarily data-driven: New big data
We can measure things we could not measure before
Analytics 1.0 Traditional Analytics
Analytics 2.0 Big Data
Analytics 3.0 Fast Business Impact for the Data Economy
, Digging in the Goldmine…
Just because you can measure it does not make it useful
à What decision will you make based on this (new) data?
Analytics 3.0 will have the decision at the center of it all!
- A seamless blend of traditional analytics and big data
- Analytics integral to running the business; strategic asset
- Rapid an agile insight delivery
- Analytical tools available at point of decision
- Cultural evolution embeds analytics into decision and operational processes
MANAGING BIG DATA AND BUSINESS ANALYTICS IS DIFFERENT FROM REGULAR IT
INITIATIVES
1. Big Data has several unique challenges, as well as new opportunities
– Four V’s
2. Big Data and Analytics projects cannot be managed like regular software projects
– Much more like R&D than building an application
3. Analytics (or data science) is a very different way of thinking about data than normal statistics
– Decision-Oriented instead of Model-Oriented
4. Data-analytic thinking can become so integrated in a company (and in several leading companies it
already is!) that it transforms the company
– Analytic-driven companies do better!
Terminology
• Data strategy refers to the strategy for creating and capturing value from data (either existing
or new)
– Requires Data-Analytic Thinking: being able to assess whether and how data
can improve performance
• Data science involves principles, processes, and techniques for understanding phenomena
via the (automated) analysis of data.
• Data mining is the extraction of knowledge from data, via technologies that incorporate
the principles from data science
