Data Analytics as a Service - Welkom op de site van de Atos

Transcription

Data Analytics as a Service - Welkom op de site van de Atos
Data Analytics as a Service
David van Steen – Product Manager
Marcel van den Bosch – Data Scientist
3-Feb-15
2
Data Analytics as a Service
Agenda
What is Data Analytics?
Data Analytics Do-It-Yourself
DAaaS platform
Join Data Analytics @ Atos
3
Data Analytics as a Service
What is Data Analytics?
Pattern Recognition
▶ Collecting, organizing and analyzing
large sets of data to discover patterns
and other useful relations
Text Mining
▶ Academic discipline for 25+ years
Network Analysis
Simulation
Advanced
Visualisation
Optimisation
▶ New processing capabilities enable for
a powerful extension of usable
applications
▶ Part of the Information element of
Gartner’s “Nexus of Forces”
4
Atos ambition
Vision, strategy and 2016 ambition presented at Atos investor
day, November 15th, 2013
Become a Tier 1 and THE preferred European global IT brand
Growth through new offerings
Application
Management
€ 41bn
Merchant
Services
€ 30bn
Work
place /
BYOD
€ 25bn
Cloud
Services
€ 50bn
IT
Consolidation
€ 47bn
MES/PLM
€9
bn
ECM &
Collaboration
€
15bn
Smart
Utilities
Digital
Security
€ 21bn
5%
7.5%
Big Data /
Advanced
Analytics
€ 36bn
Mobility
Services
€3
bn
€7
bn
25%
10%
2013 market
size
50%
Market CAGR over the 2014-2016 period*
5
Data Analytics as a Service
Agenda
What is Data Analytics?
Data Analytics Do-It-Yourself
DAaaS in action
Join Data Analytics @ Atos
6
Data Analytics: Doe-het-zelf
Fraude detectie door de RDW
▶ Onderzoeksvraag:
– Hoe kun je achterhalen wanneer er sprake is van tellerstandfraude?
▶ Achtergrond:
– RDW heeft per 1 januari 2014 de registratie en de database van Stichting
Nationale Auto Pas (NAP) overgenomen.
– Per 1 januari 2014 is het terugdraaien van de tellerstanden verboden.
– Issue: Voor naar schatting 5% van de voertuigen is de tellerstand
gemanipuleerd. Bij import auto’s ligt dit percentage nog veel hoger.
▶ Focus:
– Per regio
– Per autotype
– Per autobedrijf
– Per eigenaar
7
Data Analytics as a Service
Agenda
What is Data Analytics?
Data Analytics Do-It-Yourself
DAaaS in action
Join Data Analytics @ Atos
8
Vitens Use Case
Context
▶ Drinking water company
Vitens Innovation Playground
▶ Production, purification and distribution
of drinking water
–
–
–
–
–
Customers: 5,5 million
Turnover: 352,4 mEuro
Network: 49.000 km
Production locations: 96
Employees: 1402
-
Dedicated part of the distributionnetwork for test and validation of new
technologies
-
600 km2, 2000 km piping
-
R&D subjects: Smart-meters, Water
quality, Predictive analytics, Asset
optimization
9
Vitens Use Case
Proof of Concept
Data Analytics as a Service
▶ Use data analytics to predict leakages
in the drinking water distribution
system
Visualization
Analytic Apps
▶ Scope:
– Vitens Innovation Playground
– 26 locations, 161 sensors
– 1 year of data
XHQ
Connector
&
Cloud
Gateway
Cloud Data Analytics
▶ Goal:
Cloud Data Cloud Data
Management Sources
▶ Partners:
– KWR
– Hydrologic
€
– Develop predictive algorithm
– Show integration with current IT
solutions using standard algorithm
•
•
•
•
Joined Siemens & Atos development
Cloud based data analytics platform
Siemens XHQ connectivity to datasources
Big data enabled solution
10
Vitens – DAaaS Demo
Data in Osisoft PI
Dag 1
Dag 2
Lekkage
11
Vitens – DAaaS Demo
Details, Location & Sensors
12
Vitens Use Case
Results
▶ Predictive Algorithm
– Approach defined, preliminary model
build and next steps defined
– Predictive patterns found in subset of
the data
Vitens Data
DAaaS platform
▶ Integration
– PoC shows complete DAaaS workflow
– Vitens dataset used in demo Osisoft PI
– Standard algorithm applied
– Geo-visualization of results within
DAaaS portal build
▶ Next steps:
– Improve algorithm and extend dataset
– Connect to real-time data
XHQ gateway
GEO Visualization
13
Predictive Maintenance Use Case
Case description
Short time period
Hours/days
Low
Downtime involved
Time to React
Cost impact
14
Extended downtime
Minutes / seconds
High
Predictive Maintenance Use Case
Case description
Data from manufacturing lab and specifically a bearing dataset.
Test Rig Setup (from Qiu et al):
Four bearings were installed on a shaft. All bearings are force lubricated.
Rotation speed was kept constant at 2000 RPM by an AC motor coupled to the
shaft via rub belts. A radial load of 6000 lbs is applied onto the shaft and
bearing by a spring mechanism.
15
Predictive Maintenance Use Case
Analysis flowchart
DATA COLLECTION
EXTRACT USEFUL
FEATURES FROM THE
DATA
GET
VISUAL/STATISTICAL
PROPERTIES
16
Predictive Maintenance Use Case
Analysis flowchart
1
4
GET THE MOST USEFUL
FEATURES
K-MEANS
CLUSTERING
ARTIFICIAL NEURAL
NETWORK
▶ Getting most useful features
– Go from 46 features to the 14 most useful ones.
▶ K-Means clustering
– See if we can group the failure and suspect states (limited accuracy)
▶ Artificial Neural Network
– Prediction model with 92% accuracy
17
Results
▶ An Artificial Neural Network (ANN) is best suited for predicting a correct Suspect
state leading to a Failure state, with an accuracy of 92%.
▶ Fast-Fourier Transform frequency analysis and feature analysis (domain,
statistical & visual) provide excellent early warning.
▶ One bearing gave us a 12 day early warning.
▶ DAaaS has provided us with an analytical environment that is at least
15x faster than a local environment.
▶ This methodology is a fundamental building block for predictive cases in
different markets, using Machine Learning – Clustering and Classification.
18
Data Analytics as a Service
Agenda
What is Data Analytics?
Data Analytics Do-It-Yourself
DAaaS in action
Join Data Analytics @ Atos
19
Data Analytics @ Atos
▶ Blue Kiwi spaces:
– Big Data Analytics BTN
– Industrial Data Analytics (IDA)
▶ White Papers:
– Data Analytics as a Service
▶ Typical training:
– Certified Big data scientist
– Solutions: Hortonworks, Pivotal, Cloudera
– Data science:
• Introduction to R and CRAN libraries, SWIRL
• Coursera courses
▶ Contact us!
20
Data Analytics as a Service
Questions
21
Thank you
Atos, the Atos logo, Atos Consulting, Atos Worldline, Atos Sphere,
Atos Cloud and Atos WorldGrid
are registered trademarks of Atos SA. June 2011
© 2011 Atos. Confidential information owned by Atos, to be used by
the recipient only. This document, or any part of it, may not be
reproduced, copied, circulated and/or distributed nor quoted without
prior written approval from Atos.
3-Feb-15