analytics in your enterprise
TRANSCRIPT
• organizations have more data than ever at their disposal.
• actually deriving meaningful insights from that data—and converting knowledge into action—is easier said than done.
• There’s no single technology that encompasses big data analytics.
• several types of technology work together to help Organization get the most value from Their information.
Big Data Analytics
Real-World Applications
o Portfolio analysis and to predict the
impact of global events on financial markets.
Customer experience management and network capacity planning and optimization.
Music recommendations based on user data.
predict what the customer wants to see before he or she knows what they want!
Song identifications and predict the popular artists and genres that will get attention in the upcoming years.
Monitor financial market activities and catch illegal insider trading activities in the financial markets.
Track patient signs using sensor data.
Reduce their claims cost through better fraud detection.
Detect and prevent cyber-attacks and criminal activity.
Predict trends and lay down preparation plans to meet future demand.
Measure player efficiency and defensive effectiveness.
Source - http://upxacademy.com/2016/05/31/big-data-and-analytics-use-cases-in-8-industries/
• a single platform to address all analytics styles
• We deliver: • Batch Analytics
• Real time Analytics
• Interactive Analytics
• Predictive Analytics
• WSO2 Analytics Platform uniquely combines the above styles to turn data from IoT, mobile and Web apps into actionable insights.
WSO2 Analytics Platform
• high-level, SQL query-like languages
• Client Applications are agnostic of the Analytics Components • Common set of receivers/publishers for all analytics types
• Common format for events
• Leverage leading open source projects e.g. Storm and Spark and contribute back (such as Siddhi).
Analytics Strategy
• Open Source
• Rich, extensible, SQL-like configuration language
• Rich set of data connectors, which can be easily extended
• Events only need to be published once from applications to the platform, and can be consumed by batch or real time pipeline.
• Part of the overall WSO2 platform
Key Differentiators
AgentHolder. setConfigPath (getDataAgentConfigPath ());
DataPublisher dataPublisher = new DataPublisher(url, username, password);
String streamId = DataBridgeCommonsUtils.generateStreamId(HTTPD_LOG_STREAM, VERSION);
Event event = new Event(streamId, System.currentTimeMillis(), new Object[]{"external"}, null, new Object[]{aLog});
dataPublisher.publish(event);
Collecting Data: Example
Initialize the data publisher
Generate the stream ID for the stream to which the event will be published
Create and Publish Event
As a prerequisite, the streams must be defined in the receiver server (WSO2 DAS/CEP)
• Events are the lifeline of WSO2 CEP/DAS.
• They not only process data as events, but also interact with external systems using events.
• An Event is a unit of data
• an event stream is a sequence of events of a particular type.
• The type of events can be defined as an event stream definition.
Events , Streams and Event
Stream Definitions
Batch Analytics
Generating insight by processing large amounts of stored data ● KPI Statistics
○ Application Statistics Monitoring
○ Network / Service Statistics ○ Sensor Data Aggregation
● Solving Optimization Problems ○ Urban Planning ○ Revenue Distribution Analysis
Source: www.e-
deal.com
• Batch analytics reads data from a disk (or some other storage) and process them record by record
• “MapReduce” is the most widely used technology for batch analytics
- Apache Hadoop - Apache Spark 30X faster and much more flexible
• Analytics (Min, Max, average, correlation, histograms, might join or group data in many ways)
• Key Performance indicators (KPIs) – - e.g. Profit per square feet for retail
• Presented as a Dashboard
Batch Analytics
• Powered by Apache Spark
• up to 30x higher performance than Hadoop
• script-based analytics powered by Spark SQL
• Persist Data in A Database (RDBMS/NON-RDBMS) and process Using Spark Queries and persist analyzed data in RDBMS
WSO2 Data Analytics Server
Real-time Analytics
Making sense of fast moving data
● Sports ○ Real-time Analysis of Player
Performance ○ Real-time Match Analysis
● Geo-Spatial ○ Traffic Monitoring and Alerting ○ Geo-fencing
● Finance ○ Stock Market Monitoring
● Anomaly Detection ○ Fraud Detection ○ Network Intrusion Detection ○ Server Health Monitoring
Source: www.promojam.com
• For some use cases, the value of insights degrades very quickly with time.
• We need technology that can produce outputs fast. • Static Queries, but need very fast output (Alerts, Real-time
control)
• Dynamic and Interactive Queries ( Data exploration)
Real-TIME Analytics
• WSO2 CEP facilitates • Real time event detection
• Correlation
• Notifications/alerts, visualization tools
• Siddhi - a high-performance streaming processing engine
• WSO2 CEP is configured using the Siddhi query language
• suited for complex queries involving time windows, as well as patterns and sequences detection.
• CEP queries can be changed dynamically at runtime using templates.
WSO2 Complex Event Processor
Interactive Analytics
Near Real-time Indexed Data Search
● Log Analysis ○ Application / System Logs
● Activity Monitoring ○ Tracking Message Flows
● Fraud Detection ○ Executing queries to lookup
related data in a detected fraud situation
● HL7 Data Exploration ○ ESB HL7 Transport Interfaced
with DAS Source: befoundonline.com
• Best way to explore data is by asking Ad-hoc questions
• Interactive Analytics (search) let you query the system and receive fast results (<10s)
• Shows data in context (e.g. by grouping events from the same transaction together)
• Built using Lucene based Indexes.
Interactive Analytics with WSO2
DAS
Predictive Analytics
Analyze Existing Data to Predict Future Events
● Next Value Prediction ○ Sales Forecasts ○ Electricity Loads
● Classification ○ Product Categorization ○ Customer Segmentation
● Anomaly Detection ○ Fraud Detection ○ Preventive Maintenance
● Other ○ Handwriting recognition
• Machine learning • Takes in a lot of examples, and builds a program that matches
those examples.
• Specifically, that program is said to learn from experience E with respect to some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, improves with experience E.
• We call that program a “model”
• A Lot of Machine Learning tools • R ( Statistical language)
• Sci-kit learn (Python)
• Apache Spark’s MLLIB and Apache Mahout (Java)
Predictive Analytics
• Powered by Apache Spark MLlib
• Analyze data using machine learning algorithms
• Build machine learning models
• Compare and manage generated machine learning models
• Predict using the built models
Predictive Analytics with WSO2
Machine Learner
WSO2 Solutions Based on the
Analytics Platform ● WSO2 Fraud Detection Solution
○ Built for detecting credit card fraud ○ The rules extensible with customized Siddhi execution
plans for any type of fraud detection ○ Currently uses Real-time and Interactive Analytics
features ● WSO2 Log Analytics Solution
○ Distributed indexing and searching of any type of logs stored in the system
○ Notifications support with Real-time event processing features
○ Application / Server health prediction with Machine Learning
○ Uses Interactive + Real-time Analytics + Machine Learning features
Source: www.retrospective.centeractive.com
Source: multichannelmerchant.com
Minimum HA Deployment for DAS
2 Node Deployment Use RDBMS to Store Data If need to scale Higher Use HBASe/Cassandra