9 Service Analytics

Question 1

Which phenomenon contributes to the formation of huge amounts of data, also called big data?

Answer 1

Internet of Things -> every electronic object generates data

Industry 4.0 -> A lot more data gets “harvested” in all industries

Question 2

what are the four “Vs” of big data?

Answer 2

variety
volume
velocity
veracity (Wahrhaftigkeit)

Question 3

Why is Big Data such a big thing?

Answer 3

People expect new insights by analysing this data
- find structures

-> these new insights shall enable new business models or bring to light possibilities to makes processes more efficient

Question 4

Name the two levels big data will be exploited and name examples

Answer 4

Centralized level: Central data model and intelligence
- transportation systems, warehouses

Autonomous/ decentralized intelligence

• Self-managing traffic lights
• packages will find their way automatically

Question 5

Name 8  (Business) Analytics methods 
\+ group them in "basic" and "advanced"
\+ sort them according to their degree of intelligence/ competitive advantage

Answer 5

Basic:
Standard reporting
Ad hoc reporting
Drill down
Alerts

Advanced: 
Forecasting
Simulation
Predictive modeling
Optimization

Question 6

Name three kinds of analytics

Answer 6

Descriptive
Predictive
Prescriptive

Question 7

In which departments of a company are more likely to rely on data and analytics?

Answer 7

with increasing value:

Customer
HR
Strategic
Operational
Financial

Question 8

Distinguish supervised from unsupervised machine learning

Answer 8

supervised:

• we need to train the model
• > set of know problems/ answers needed to train
• typical tasks: regression or classification
• Example: Tell a child to sort cars into sports cars and SUV after telling him what their characteristics are

unsupervised:

• identify previously unknown patterns
• > outcome might be a structure we haven’t been thinking of yet
• typical: clustering/ association rules
• Example: Tell a child to sort cars but don’t give any criteria

Question 9

Definition: Machine learning

Answer 9

algorithms that can learn from and make predictions on

data

Question 10

Definition: Data Mining/ Knowledge discovery in databases

Answer 10

The nontrivial process of identifying valid, novel, potentially useful, and understandable patterns in
data

Question 11

Draw the Crisp- DM circle

Answer 11

Business understanding Data understanding

Data Understanding -> Data preparation

Data preparation Modeling

Modeling -> Evaluation -> Deployment

Evaluation -> Business understanding

Question 12

Describe the term Data Science using a Venn Diagram

Answer 12

Computer Science + Statistics = Art zone, solving problems that never appear

Computer Science + Business Application = Danger zone, automation of gut feeling

Statistics + Business Application = Theory zone, solution without implementation

All three disciplines are needed!

Question 13

Explain shortly multivariate regression

is it used for supervised or unsupervised learning?

Answer 13

predict the value of a response variable with help of its correlation with other variables

• find the right few (~ 2-4) variables that explain the response variable realistically
• not too many -> overfitting

-> used for supervised learning

Question 14

describe the method “classification” shortly and name techniques for it

Answer 14

One variable indicates class membership
The other variables are used to predict it

Techniques:
Naive Bayes
k-Nearest Neighbor
Decision Trees (Quinlan’s ID3, C4.5)
Logistic Regression
Neural Networks

Question 15

Name two applications of unsupervised learning

Answer 15

Clustering/ Segmentation

Association rules

Question 16

Explain Clustering shortly

Answer 16

Arrange data objects so that nearby objects are similar

• > Algorithm derives (herleiten) the pattern
• K- Means algorithm to cluster n objects into k groups

Question 17

Explain shortly association rules

Answer 17

Algorithm should find rules/ dependencies in a set of objects

Example:

• When someone buys beer, how likely is it that he also buys chips (beers -> chips)
• When someone buys chips, how likely is it that he also buys beer (chips -> beer)
• How often does someone put a specific object in his cart?

Question 18

What kind of algorithms exist for text mining? Name examples for applications

Answer 18

Classification
Clustering
Language analysis (what’s the meaning, but also what kind of person is that)

Applications:

• Identifying personal traits from texts -> create a personality portrait
• capture information about what customers want from social networks