PROČ UŽ SE NEOBEJDETE BEZ ANALÝZY DAT JAKUB CHOVANEC - IDG KONFERENCE 3.6.2015
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
KDO JSME
#1 v poskytování datové analytiky a služeb v oblasti Business Analytics a Business Intelligence 39 let na trhu 16 let v České republice
CO DĚLÁME
Integrovaný soubor softwarových produktů a služeb Informační management Analytika Business Intelligence
ZKUŠENOSTI
NAŠI ZÁKAZNÍCI
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
70,000 instalací ve 139 zemích po celém světě 23% z výnosů investovaných do vývoje $3,09 mld. ve výnosech v roce 2014 91 firem umístěných v top 100 v žebříčku Fortune Global 500 Klienti ze soukromého i veřejného sektoru po celém světě V ČR: Česká spořitelna, Česká pojišťovna, Kooperativa, Raiffeisenbank, Allianz, GE Money, ČPP, T-Mobile, DIRECT Pojišťovna, ČSÚ, Komerční banka, VIG, Kooperativa, PČS, Česká pošta, Ministerstvo zemědělství
DATA GENERATED AROUND US
Risk
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
Commerce E-commerce CRM Telco
Fraud
Utilities
Manufacturing
RISK WILL THEY PAY THE DEBT?
Risk
•
Credit risk scoring •
•
Internal & external data – social networks
User behavior on website •
„smart forms“
- position - keywords in text - average time on position
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
MARKETING WHAT IS THE CUSTOMER BEHAVIOUR?
•
Customer 360 view •
Propensity to buy • Next best offer • Churn analysis
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
Commerce E-commerce CRM
FRAUD WHO AND HOW? HOW TO FIND OUT?
•
Insurance fraud • • •
•
Rules Social Network Analysis Text Mining
60+ millions CZK saved in ½ year
Banking - Online fraud • • • •
Fast data – Event Stream Processing Analyzes transactions as they are received Alert Investigation
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
Fraud
ONLINE FRAUD THE CLIENT’S PROBLEM?
Fraud
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
UTILITIES HOW TO MANAGE RISK?
Utilities
• •
Predictive maintenance Smart meter analysis
Real-time analytics
ALERT VISUALIZATION REPORTS DASHBOARDS INSIGHTS
Batch analytics
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
MANUFACTURING HOW TO USE DATA FOR QUALITY MANAGEMENT?
•
Quality management •
Parts fitting •
Sensor data analysis
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
Manufacturing
MANUFACTURING HOW TO USE DATA FOR QUALITY MANAGEMENT?
•
Quality management •
Parts fitting •
Sensor data analysis
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
Manufacturing
MANUFACTURING HOW TO USE DATA FOR QUALITY MANAGEMENT?
•
Quality management •
Parts fitting •
Sensor data analysis
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
Manufacturing
TELCO – BIG DATA
1. Activation
2. Set Preferences
9. Use in store
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
3. See Featured Offers
8. Accept
4. Receive push notification of local offer
5. View Offer on Map
7. Get Offer
6. View Offer Details
Commerce & Telco
WHY DO WE NEED TO ANALYZE DATA?
INSIGHT
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
MONEY
HOW DATA VISUALIZATION
•
Goal: easy to understand graphical representation of data using • • •
• •
Shapes Colors Color intensity Position of shapes …
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
HOW DATA VISUALIZATION – PATTERNS, RELATIONSHIPS, …
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
SAS VISUAL EXPLORE YOUR DATA ANALYTICS •
Visualization types • • • •
•
Table analysis (pivot) Common graph types Powerful filtrations, selections Geo-analysis
Analytics for business people • • • • •
Forecasting What-if analysis Correlation matrix Heat, tree maps Text analytics
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
SAS VISUAL RAPID MODEL PROTOTYPING STATISTICS
•
Exploratory predictive modeling, in-memory • • • • •
• •
Linear regression Logistic regression GLM Clustering Model comparison
Score code Group by modeling
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
SAS VISUAL GROUP BY MODELING STATISTICS
• • •
Big Data – one model doesn´t fit all Find „groups“ (clustering) Make models for each „group“ • • •
Market segment Customer segment …
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
HOW IS SAS LEVERAGING DEPLOYMENT PATTERNS HADOOP? Traditional Approach
In-Memory Approach
Memory
Data
Data
SAS
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
SAS
In-Database Approach
Data
SAS
SAS&HADOOP DISTRIBUTIONS
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
PREFERRED CLOUDERA & HORTONWORKS
WHY TO MAKE RELEVANT DECISIONS!
INSIGHT
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
MONEY
THANK YOU!
Cop yrig ht © 2015, SAS Institute Inc. All rig hts reserv ed.
