Los datos hablan. El impacto de Big Data en la Sociedad y la Economía
Dr Pedro A. de AlarcónSenior Data Scientist. External Positioning and Big Data for Social Good
@LUCA_D3
1. Let Data Speak
2. ¿Qué está haciendo Telefónica?
3. Algunas ideas para durante y después del Máster
4. … y un buen ejemplo.
@LUCA_D3
@LUCA_D3@pdealarcon
¿Qué está pasando?... Adapt or Die
@LUCA_D3
@LUCA_D3
Let Data Speak01
Are you listening?
@LUCA_D3
Using CDR traffic to understand the impact of an earthquake in Mexico
Are you listening?
@LUCA_D3
Using sports analytics software to understand footballer behaviour
Xavi: movements faster than 5 m/sec
Messi: movements faster
than 5 m/sec
Time at each position sec.
Big Data for Football Analytics
@LUCA_D3
Time at each position sec.
Open Data for Customer Benefit
@LUCA_D3
Using open data to provide automated Oyster refunds in London
Open Data for Marketing and Brand Awareness
@LUCA_D3
Using open data as an input for a real-time advertising campaign in London
Big Data is no longer an emerging technology…
@LUCA_D3
“We’ve retired the Big Data hype cycle. I know some clients may be really surprised by that because Big Data was a really important one for many years.” Betsy Burton, Gartner
Tools and Infrastructure
@LUCA_D3
Difficulty
Val
ue
Descriptive Analytics
Diagnostic Analytics
Predictive Analytics
Prescriptive Analytics
What happened?
Why did it happen?
What will happen?
How can we make it happen?
Source, Gartner 2012
Volume Velocity
Variety
3 Vs of Big
DataTerabytesRecords
TransactionsTables / Files
BatchReal-timeStreams
Near-time
StructuredUnstructured
Semi-structuredAll of the above
Data Science
@LUCA_D3
“Data Scientist is the sexiest job in the 21st century.”
“Big Data expertise demand has increased by 85% in the last year.”
“21.000 Data Engineer jobs posted online.”
“90.000 Data Engineer jobs posted online.”
@LUCA_D3
¿Qué está haciendoTelefónica?02
We have launched LUCA Data-Driven Decisions
@LUCA_D3
Our goal: to empower other organizations to become more data-driven
The Big Data journey has different maturity stages
@LUCA_D3
• Interested in understanding potential of Big Data
• Keen to run POCs
• Lack of internal know-how and capabilities
• Data sources unknown / uncontrolled
• Strategy and uses cases not defined
Exploration
Data-Driven
• Big Data already implemented in a “data-centric” organization
• Data is intrinsic in decision-making
• Data brings organization closer to its customers
• In-house expertise in data science and data engineering
• External data sources used regularly to generate richer insights
• New data-driven business models and processes in creation
Transformation
• Decision made to leverage Big Data
• Require methodology and expert resources
• Need technology and infrastructure
• Analytics capabilities to be deployed / outsourced
• End-to-end transformation programme planned, including cultural and org. changes
And this is Telefonica’s Big Data journey – so far
@LUCA_D3
2011 20132012
Explorationproving Big Data’s
potential
Smart Steps launched
SNA pilots marketing and churn
Global BI organization created
Data to become an asset
20152014
Transformationaccelerating adoption
Smart Digits launched
China Unicom Joint Venture
Smart Steps International roll out
Key use cases executed
Strategic data sourcing installed
20172016
Data-Drivendemocratizing Big Data
Big Data B2B unit launch
Niche Big Data consultancy acquired
Cultural transformation
CDO (Chema Alonso) appointed
Breaking all kinds of silos
Big Data roadmap for each Operating Business
New data-driven P&S
These are the components to work on during the journey
@LUCA_D3
Data
Engineering
Tools &
Infrastructure
Data
Science
Business Insights
Strategy and Transformation (Strategic Assessments, Corporate Transformation Programmes, Culture and Talent programmes)
End 2 End Security Wrap
Data Engineering – more data sources
@LUCA_D3
Big DataData Warehouse
+ External DataGeolocation
Call Centre Content
Channels (online & offline)
TV Audiences
Q of E
Sociodemographic
Products and Services
Consumption
Customer
Web LogsDescriptive
WHO ?
Behavioral WHAT, WHERE ?
InteractionHOW ?
AttitudinalWHY ?
All of which is applied to the Business
@LUCA_D3
Product and Innovation
Operational and Finance
Channels
Sales and Marketing
Network and Infrastructure
Business optimisation
which have allowed us to create value internally:
@LUCA_D3
Smart Marketing
Churn Reduction & Optimisation
Network Planning
Optimization
Real Time contextual campaigns
Video analytics
AData-Driven
Online channel
E2E churn approach
Multi-channel analytics
Fibre/LTE deployment
Supply/demand analysis
Value based analysis
Procurement analytics
Fleet management optimisation
Business Insights for Sustainable transport
@LUCA_D3
Using mobile data to measure the sustainability of commuting in Barcelona
Business Insights for Tourism Insights
@LUCA_D3
Using mobile data to understand the catchment and profile of tourists to the Fallas Festival
Business Insights for Tourism Insights
@LUCA_D3
Using mobile data to understand the catchment and profile of tourists to Spain
Business Insights for Smarter Communication
@LUCA_D3
Using analytics to optimize global communications in multinationals
Big Data for Social Good
@LUCA_D3
Using CDR traffic to understand the impact of floods in Mexico
@LUCA_D3
Algunas ideas03
Aprovechad las clases presenciales
@LUCA_D3
Vs.
Leed blogs de Ciencia de Datos
@LUCA_D3
… os ayudarán a averiguar cuales son vuestrosverdaderos intereses
Expresad vuestros intereses y mostrad vuestro valordiferencial ANTES del primer trabajo… o como escribirun CV aplastante sin usar Word.
@LUCA_D3
• Kaggle• Github• Ipython Notebooks• Hackatons• Open data Apps (Code4Good)• StackOverflow• Blog• Twitter & Linkedin
Los datos deben contar una historia: contadla bien.
@LUCA_D3
http://www.forbes.com/sites/brentdykes/2016/03/31/data-storytelling-the-essential-data-science-skill-everyone-needs/#1098ffadf0c8
No olvideis las herramientas comerciales
@LUCA_D3
Cambio de enfoque
@LUCA_D3
Teoría Práctica
ValidaciónCaso de uso/ Problema
Preparaciónde datos Teoría
ProgramaciónModelos /
Visualización
Industrialización
Máster
Empresa
Si Mahoma no va a la montaña…
@LUCA_D3
• Lenguajes & Machine Learning: Python, R, Scala
• Big Data (básico): Spark, Hive.
• Demandas no bien cubiertas: Modelos Predicitivos en Real Time.Visualizaciones impresionantes (D3)Mapas (PostGis)
• Vienen fuerte: Deep LearningTensor FlowAI+NLP (Chatbots)
@LUCA_D3
Un buen ejemplo ;-)04
@LUCA_D3
QUE ESPERABA…
• Conocer estructura de una gran empresa
@LUCA_D3
QUE ESPERABA…
• Tratar con datos reales y muy diferentes
@LUCA_D3
QUE ESPERABA…• Grandes proyectos
@LUCA_D3
QUE ME HE ENCONTRADO…
• Una gran empresa
@LUCA_D3
QUE ME HE ENCONTRADO…• Aprendizaje teórico continuo
SERIES TEMPORALES
ANÁLISIS DE REDES
ESTADÍSTICA
PREPROCESAMIENTO
@LUCA_D3
QUE ME HE ENCONTRADO…
• Aprendizaje tecnológico continuo
@LUCA_D3
EJEMPLOS …
• GLOBAL RIDER
https://www.youtube.com/watch?v=XeCTZcHqqyc
• ELECCIONES ESTADOS UNIDOS
http://data-speaks.luca-d3.com/2016/11/big-data-and-elections-we-shine-light.html
Global Rider
@LUCA_D3
https://www.youtube.com/watch?v=XeCTZcHqqyc
@LUCA_D3
Elecciones USA 2016 & Twitter
http://data-speaks.luca-d3.com/2016/11/big-data-and-elections-we-shine-light.html
TwitterSinfonier(11Paths) MongoDB
PhilipsHue + Spotfire
RealTime Filter Python
Elastic + Kibana
vs
@LUCA_D3
Elecciones USA 2016 & Twitter
http://data-speaks.luca-d3.com/2016/11/big-data-and-elections-we-shine-light.html
@LUCA_D3
Let your data speak
Thanks for listening!@vrbenjamins / @LUCA_D3
www.luca-d3.com