NOVA

IMS 1
Information
Management
School
INTRODUCTION
TO
DATA SCIENCE

Data Science for Marketing

© 2021-2024 Nuno António
Acreditações e Certificações
Instituto Superior de Estatística e Gestão da Informação
Universidade Nova de Lisboa
 Summary
1. Introduction
2. Tasks in Data Science
3.Applications in Marketing

2
 1.1
0

Introduction
Introduction to Data Science

3
 What is Data Science
Data science involves principles, processes, and techniques for
understanding phenomena via the (automated) analysis of data.
§Interdisciplinary field
§Relates to Data Mining, Machine Learning, Network theory, and
Big Data
§Uses scientific methods, processes, algorithms and systems to
extract knowledge and insights from structured and unstructured
data
§Applies knowledge and actionable insights from data across a
broad range of application domains

4
 Insights
§Identify new, relevant, and non-trivial
information
§Most focus on understanding consumer
behavior
§Must quantify causality
§Must provide a competitive advantage
§Must generate financial implications
§Ultimately, is about action-ability

5
 Data-driven decision-making
Data Science is the base for Data-Driven Decision-
making (DDD). DDD refers to the practice of basing
decisions on the analysis of data, rather than purely on
intuition.
According to Brynjolfsson, Hitt, & Kim (2011) in
“Strength in Numbers: How Does Data-Driven
Decision-making Affect Firm Performance?” firms who
apply DDD have an output and performance that is 5-
6% higher than other firms

How should a marketer select advertisements:

§ Based on his/her long experience in the field and eye
for what will work, or
§ Based on the analysis of data of how consumers
react to the different type of ads [Provost & Fawcett 2013]
6
The intelligence hierarchy

Prescriptive
POWER OF Modeling
INFORMATION

Adapted from Cokins (2012)

Predictive
Modeling
Descriptive
Ad Hoc Modeling
Reports
Standard & OLAP
Raw Reports
Data

ROI (€)
DATA INFORMATION KNOWLEDGE INTELLIGENCE

7
 What is today known as “Analytics”

DESCRIPTIVE PREDICTIVE PRESCRIPTIVE

QUESTIONS

What happened? What will happen? What should I do?

What is happening? Why will it happen? Why should I do it?

Business reporting Machine Learning Optimization

ENABLERS

Dashboards Text mining Simulation

Scorecards Web/media mining Decision modeling
Data warehousing Forecasting Expert systems

Well-defined business Accurate projections of

OUTCOMES

Best possible business

problems and future events and
decisions and actions
opportunities outcomes

8
Marketing Analytics

what
will happen
what
happened why
will it happen
what what
is happening should we do
why
should we do it
9
 Data Science combines
§ Data-driven approach of
statistical data analysis
§ The computational power and
programming acumen of
computer science
§ Domain-specific expertise and
business intelligence

[http://drewconway.com/zia/2013/3/26/the-data-science-venn-diagram] 10
11
 But remember…
You do not need to know
everything by memory nor to be
an expert in every analytical tool.
You need to know:
§How to make questions and to
whom
§What techniques and
algorithms apply to the
business problem
§What tools to use
§Search for information and
interpret the documentation

12
 1.2
0

Tasks in Data Science

Introduction to Data Science

13
 Main goal in Data Science
Like in Data Mining, the main goal of Data Science is to uncover
interesting patterns in data. To be interesting, a pattern must:
1. Easily understood by humans
2. Valid on new or test data with some degree of certainty
3.Potentially useful
4.Novel
or, validate a hypotheses the user is sought to confirm

14
 Types of patterns
Additional information
covered in Descriptive
Analytics in Marketing, Networks covered in
Business Intelligence I/II, Social Network Analysis,
Big Data for Marketing, Social Media Analytics,
Data characterization
among other Class/Concept description among other
Data discrimination

Frequent patterns, Frequent patterns

Associations, and Correlations Association analysis

Classification
PATTERNS Predictive Analysis
Regression

Cluster Analysis

Covered in Machine
Outlier/Anomaly Analysis
Learning for Marketing,
Big Data for Marketing,
Text Mining, among
other
15
 Class/Concept descriptions
Useful to describe classes or concepts (e.g., segments of customers)
in summarized, concise, and yet precise terms
Data characterization Data discrimination
§ Summarization of the general § Comparison of the general
characteristics of a target class of features of the target class data
data (for example the objects against the general
characteristics of products with features of objects from one or
sales that increased by 10% in the multiple contrasting classes (e.g.,
previous year) products with sales that
§ Data obtained from a increased by 10% in the previous
transactional or OLAP database year against those that decreased
30%)
§ The output is usually pie charts, § The type of data and outputs are
bar charts, line charts, pivot
tables, or crosstabs similar to data characterization

16
 Frequent patterns, associations, and correlations

Frequent pattern mining investigates recurring relationships in a

dataset, like the co-occurrence of two or more objects of interest.

Types of frequent patterns:

§Frequent itemsets: typically refers to a set of items that often
appear together in a transactional dataset (e.g., milk and bread in a
grocery store dataset – “Market basket analysis”)
§Sequential pattern: refers to patterns that occur across time or
positions in a dataset (e.g., a customer first buys a laptop, followed
by a webcam, and then an external hard disk)
§Structured pattern: structural patterns (graphs, trees, or lattices)
that can be combined with itemsets or sequential patterns

17
 Frequent patterns, associations, and correlations

Frequent structured

source: https://neo4j.com/blog/analyzing-panama-papers-neo4j/
patterns example

“The Panama Papers”

– The International
Consortium of
Investigative
Journalists Exposed
highly connected
networks of offshore
tax structures

18
 1.3
0

Applications in Marketing
Introduction to Data Science

19
 Customer-oriented
§Targeting current customers
§ Segmentation based on touchpoint engagement
§ Segmentation based on purchase patterns
§ Micro-segmentation/personalization
§Finding new customers
§ Lead targeting
§ Lead scoring
§Retaining customers
§ Churn prediction
§Predicting sales
§ Demand forecast
20
 Product-oriented
§ Understanding markets
§ Understanding customers’ likes and dislikes
§ Positioning products
§ Budget optimization
§ Developing new products
§ Real-time experimentation
§ Promoting products
§ Optimize campaigns
§ Recommending products
§ Market basket analysis
§ Assessing brands and prices
§ Pricing analysis
§ Competitor's analysis
§ Predicting sales
§ Demand forecasting

21
 Algorithmic marketing
The advancement of digital marketing channels changed the game
and created an environment that requires millions of micro-
decisions to be made, which simply cannot be done efficiently
without intelligent marketing software and algorithms:
§Targeted sales promotions
§Dynamic pricing in brick-and-mortar and online stores
§E-Commerce search and recommendations services
§Online advertising

22
Data Science for Marketing
© 2021-2024 Nuno António (Rev. 2024-08-20)
Acreditações e Certificações
Instituto Superior de Estatística e Gestão da Informação
Universidade Nova de Lisboa