This document discusses big data and data science challenges and opportunities. It provides background on the author, Jose Quesada, and outlines five key challenges companies face: 1) obtaining data from end users, 2) creating a data-driven culture, 3) finding data talent, 4) breaking down data silos within companies, and 5) addressing hype around big data. The document then provides three opportunities for companies: 1) measuring their data maturity, 2) identifying the value they want from data, and 3) finding stakeholders within the company who would benefit most from increased data use. Throughout, the author advocates starting small with available data rather than waiting for "big data" to extract business value.
Memorándum de Entendimiento (MoU) entre Codelco y SQM
Big data & data science challenges and opportunities
1. Big Data & Data Science -
Challenges and
Opportunities
Jose Quesada, Phd
Director
@quesada, @dataScienceRetreat
2. Personal Background
• PhD in Machine learning, researcher at top labs
• Solving data problems for the last 15 years
• Consultant on ‘customer lifetime value’
• Data scientist at GetYourGuide
• Today, Director at Data Science Retreat
5. “Companies that have embraced a
data-driven culture—rating
themselves substantially ahead of
their peers in their use of data—are
three times more likely to rate
themselves as substantially ahead
of their peers in financial
performance” --The Economist Intelligence Unit
x3
7. "Many of my clients are clearly aware of the
importance of data, But they don't know where to
start in terms of where they should focus to get the
most value, as well as how to translate the data into
actionable insight."
Jerry O'Dwyer, a principal at Deloitte Consulting
http://www.cio.com/article/2387460/business-
intelligence/data-driven-companies-outperform-competitors-
financially.html
8.
9. Data Science Retreat mission
“Making sure we
(EU) don’t fall
hopelessly behind
the US when it
comes to
technology”
16. Bad Example: Window maker
• Real company in DE (name omitted)
• No information about what their customers care about
• No brand recognition by customers
• Exposed to cheaper competitor entering the market any time
18. Bad Example: textbook publisher
• Real companies (everywhere)
• No idea how long it takes for their customer to consume each
page of the textbook
• No information about what their customers care about
• No brand recognition by customers
• Exposed to cheaper competitor entering the market any time
27. You don’t need to have big data to
extract value from it. You can make
better decisions with your data today.
Certainly, you don’t need a Hadoop
cluster to start!
29. 1: Measure your company’s data
maturity
"When was the last time you had to defend forecasts against
actuals?“
Identify where you are on the Drake scale for data maturity.
Aim to move your company one level up
30. The Drake scale for data maturity
http://aadrake.com/the-kardashev-scale-of-data-maturity.html
Type 1
Type 2
Type 3
31. The Drake scale for data maturity
http://aadrake.com/the-kardashev-scale-of-data-maturity.html
Type 1
Type 2
Type 3
Staying out of jail.
No data roles
32. The Drake scale for data maturity
http://aadrake.com/the-kardashev-scale-of-data-maturity.html
Type 1
Type 2
Type 3
Business Intelligence,
reporting, or similar
team that may use
spreadsheets
33. The Drake scale for data maturity
http://aadrake.com/the-kardashev-scale-of-data-maturity.html
Type 1
Type 2
Type 3
Chief Data Officer or
similar role.
Reporting and ad hoc
requests previously
handled by the BI
team are now part of
a self-service
platform so any
employee can analyze
the data
34. 2: Identify what value you would like
to get out of your data
Types of value:
•Decrease risk
•Higher precision
•Foster innovation
35. 3: Identify who in the company has
the most to gain, form a coalition
Since you need to change the culture of your
company (not easy!), every stakeholder you can
recruit helps
Recruit people from outside the company if
needed
37. Data Science is a chaotic field
and people don’t really know
what they want (much less
what they need)
38. Thank You!
Check out our short courses:
Deep Learning
Scalable machine learning
Big Data Business value
---
Jose Quesada, PhD
Director, Data Science Retreat
@datascienceret
me@josequesada.com