Working as a Data Scientist in a company: big hype or big hope
1. Working as a Data Scientist in a company:
big hype or big hope
MADDALENA AMORUSO
Rome, 19 MAY 2018
2. 2
allrightsreserved
Prometeia, founded in Bologna in 1974, operates as a consultancy, software and economic
research firm. The company offers solutions in the areas of Enterprise Risk Management, Wealth
Management, Business Consulting, Asset Management and Knowledge Training. Prometeia is
based in Bologna and Milan, with additional offices in Rome, Istanbul, Moscow and London.
Company Profile
5. 5
allrightsreserved
How does the Job Market look like nowadays?
Some Key Facts, according to LinkedIn’s 2017
U.S. Top 20 Emerging Jobs Report:
› Jobs with the top growth potential are tech-
focused, with demand coming from both tech
and non-tech companies
› Machine learning engineers, data
scientists, and big data engineers rank
among the top emerging jobs
› Traditional soft skills underpin all of these
emerging jobs
› Data Scientist roles have grown up over
650% since 2012. Supply of candidates for
these roles cannot keep up with demand
› Software engineers are feeding into all of
the technology-related professions
Source: LinkedIn’s 2017 “U.S. Emerging Jobs”
7. 7
allrightsreserved
What are the problems for companies in Big Data?
Source: Gartner 2017 “How (NOT) to choose the right data and analytics service provider”
8. 8
allrightsreserved
According to Gartner report on Big Data:
› Among the interviewed, the biggest
challenge for companies is to understand
how to define and get value from Big
Data
› Also, very often companies are missing
strategies, skills and capabilities to
correctly use the data
What are the problems for companies in Big Data?
Source: Gartner 2017 “How (NOT) to choose the right data and analytics service provider”
11. 11
allrightsreserved
Why so many companies are trying to hire data scientist if:
? they do not have a clear data and analytics
strategy?
? they do not know how to get the value from their
data?
? they do not have IT tools and architectures ready to
handle these data?
Questions and answers
12. 12
allrightsreserved
Why so many companies are trying to hire data scientist if:
? they do not have a clear data and analytics
strategy?
? they do not know how to get the value from their
data?
? they do not have IT tools and architectures ready to
handle these data?
Questions and answers
…they are hiring because they want them to solve
these problems:
› they want data scientists to understand the
business needs and to define a strategic plan
› they want them to extract the best value from
the data
› they want them to setup the proper tools and
architectures to deliver
13. 13
allrightsreserved
Data Analytics Team
Data Science
› Understanding data mining and
machine learning models
(classification, clustering etc.)
› Statistical modelling
› Creativity and curiosity
› Understand performance metrics
› Feature engineering
› Visualization skills
IT Skills
› Data modelling skills (relational, non-relational)
› Data governance
› Data cleaning and preparation
› Coding skills
› DevOps / SysOp skills
› Big data skills
Business Skills
› Ask good questions
› Business and customer
understanding
› Know the constraints (e.g., legal,
ethics, market)
› Translate for non-technical
audience
› Decision making
› Storytelling
Unicorn
14. 14
allrightsreserved
Data Analytics Team
Data Science
› Understanding data mining and
machine learning models
(classification, clustering etc.)
› Statistical modelling
› Creativity and curiosity
› Understand performance metrics
› Feature engineering
› Visualization skills
IT Skills
› Data modelling skills (relational, non-relational)
› Data governance
› Data cleaning and preparation
› Coding skills
› DevOps / SysOp skills
› Big data skills
Business Skills
› Ask good questions
› Business and customer
understanding
› Know the constraints (e.g., legal,
ethics, market)
› Translate for non-technical
audience
› Decision making
› Storytelling
Unicorn
15. 15
allrightsreserved
The right team should be formed by the following roles:
› Business translators
› Data Scientist
› Data Engineers
The team should be big enough to handle several
business requests
The factors for a successful Analytics Team
Prometeia’s experience in setting up Analytics Team
The right team
Data Driven Business Plan
Having Agile Business processes
Digital transformation should be driven by proper
data driven business plans, where companies set up
objectives and key indicators to measure success
The Analytics Team should work in a flexible way,
following Agile methodologies where requirements
are subject to frequent changes driven by business
needs
16. 16
allrightsreserved
Business Value
Data Analytics Toolbox
The Data Analytics team as a toolbox to deliver value to business
The Data analytics team should be seen as a toolbox
where each tool can be used as a facilitator to
analyze use cases and create business value.
Oversimplifying, the three most important tools are:
› Techniques for structured data analysis
› Classification
› Clustering
› Regression
› Techniques for unstructured data analysis
› Text analytics
› Image recognition
› Techniques for network analysis
› Graph analysis
› Social Network Analysis (SNA)
Structured
Data
Unstructured
Data
Network
Analysis
Use
Cases
17. 17
allrightsreserved
Text Analytics Techniques
Sentiment Analysis
Natural Language Processing Topic Mining
Contenuto del topic Keywords
% di
commenti
Insoddisfazione con il
lavoro del perito
Perito, risposta,
riparazione, tempo
44.3%
Insoddisfazione con la
gestione del sinistro
Pagare, male,
informazione,
mancanza
28.6%
Difficoltà nell'avere un
contatto telefonico
Chiamare, telefono,
tempo, parlare
18.5%
Lunghi tempi di attesa per
l'assistenza stradale
Rimorchio, telefono,
minuti, aspettare,
assistenza, tempo,
auto
8.6%
Text Clustering and
Classification
Several Text Analytics techniques enable…
Text Retrieval Entity extraction
Trump wasn't in his office in Washington DC
yet when the Labor Department collected the
data used in January 2017 jobs report, so for
the sake of comparison it makes sense to
exclude the first month of the year. But in the
remaining 11 monthly jobs reports, employers
added 1.84 million jobs, according to the
December jobs report released Friday.
: Name, : Organization, : Time,
: Number, : Location
18. 18
allrightsreserved
Text Analytics in the Finance Industry
…advanced capabilities to support all the business processes of the company…
Process Automation
Data sources:
› Letters/emails/fax from customers
› Insurance medical receipts
› Agents/branches tickets
Know Your Customer
Data sources:
› Text data related to payments
› Surveys
› Feedback
› Net Promoter score
Risk Management
Data sources:
› Balance Sheet
› Contract Terms and conditions
› Claim Requests
› Transaction Data
› News Data
Operations / HR
Data sources:
› Internal historical documents
(procurement, HR)
Marketing
Data sources:
› Newspapers
› Social Network
› Blogs
› Forum
Call Center
Data sources:
› Calls to Call center (speech or text)
› Operators Notes
19. 19
allrightsreserved
Text Analytics in action – Use cases
…which are ultimately applicable to several use cases / business areas
Process Automation
Agent Tickets
› Reduction of ticket routing errors by agents
› Reduction of ticket resolution timing
› Reduction of the human resources who deal
with repetitive and simple ticket resolution
Call Center
Entity extraction from unstructured data
to enrich structured data
› Brand awareness and Social responsibility
with clients
› Extraction of information for further
developments on Pricing optimization
Process Automation
Medical Receipts
› Reduction of costs in claims handling
› Reduction of time to payment to improve
the customer experience
Process Automation
Automatic communication routing
› Reduction of costs in manual handling
› Reduction of time to direct the mail
› Reputational and operational risk mitigation
21. 21
allrightsreserved
Conclusions
› Data and algorithms knowledge is the main skill needed
in the business world nowadays
› Data-related jobs are the biggest growing industry that
spans across several verticals
22. 22
allrightsreserved
Conclusions
› Data and algorithms knowledge is the main skill needed
in the business world nowadays
› Data-related jobs are the biggest growing industry that
spans across several verticals
› Key factors for successful analytics team are: skills and
capabilities, data-driven business plan and agile
processes
23. 23
allrightsreserved
Conclusions
› Data and algorithms knowledge is the main skill needed
in the business world nowadays
› Data-related jobs are the biggest growing industry that
spans across several verticals
› Key factors for successful analytics team are: skills and
capabilities, data-driven business plan and agile
processes
› Data analytics team is a facilitator to enable use cases
and create business value
24. 24
allrightsreserved
Conclusions
› Data and algorithms knowledge is the main skill needed
in the business world nowadays
› Data-related jobs are the biggest growing industry that
spans across several verticals
› Key factors for successful analytics team are: skills and
capabilities, data-driven business plan and agile
processes
› Data analytics team is a facilitator to enable use cases
and create business value
› Now it is the time to study STEM disciplines… and
choose your path carefully.