01 Data Science for Business

Question 1

What is Big Data?

Answer 1

• Big Data is high-volumen, high velocity and/or high-variety information assets that demand cost-effective, innovative forms of information processing that enable enhanced insight, decision making and process automation.
• The three V’s: volume, velocity and variety.

Question 2

What is the general procedure of Data Science regarding extracting knowledge from data?

Answer 2

• It is assumed that extracting useful knowledge from data to solve business problems can be done systematically.
• IT is used to compile large data sets. Then data is analyzed to identify correlations to predict variables.
• Results from the analysis need to be generalizable (prevent overfitting the dataset).
• Meaningful decisions = identify the contexts in which data are created, analyzed and used.

Question 3

Describe the relation of decision making and data:

Answer 3

• Traditionally, many decisions are made based on “gut feeling” of executives.
• Nowadays, data is available at sufficient quantity and granularity to let more decisions be based on facts.
• Top-down decision making is supplemented by bottom-up data analytics.

Question 4

Why are Data Science capabilities regarded as a strategic asset?

Answer 4

• The strategic value of data science (compiling and analyzing data) can:
  
  • Transform existing opperations to be more efficient
  • Transform entire business models and generate new ways to earn profit/market share

Question 5

Give some examples of data science solutions for business problems?

Answer 5

• Classification and class probability estimations.
• Regression (estimate numerical values for an individual)
• Similarity matching (identify similar individuals)
• Clustering (group of individuals, customer segments)
• Profiling (typical behaviours of individuals/groups)
• Link prediction (connections among data items)
• Data reduction (replace large data sets with smaller datasets)
• Causal modeling (what events influence each other).

Question 6

Name some (5) approaches for data science:

Answer 6

• Statistics
• Database querying
• Data Warehousing
• Regresion analysis
• Machine learning, data mining

Question 7

Briefly describe the evolution of analytical information systems:

Answer 7

1960: MIS (Management IS) Efficiente data processing, integrated information systems, vision of automated decision making.

1970: DSS (Decision Support System) Statistical algorithms, “what if” analysis, complex and rigid structures, databases.

1980: EIS (Executive IS) Multidimensional modeling, transaction processing and decision support, top management decisions.

1990: DWH (Data Warehouse) Integration of diverse data, interactive and customized reports/OLAP, historical data.

2000: BI (Business Intelligence) KPI Systems, balanced scorecards, analytical applications, data mining.

Question 8

What are the key differences between model-driven vs data-driven decision making?

Answer 8

• Model-based decision making is based on predictions, based on statistics and operations research techniques.
• Data-driven decision making is based on providing data in one place (in a DWH).