health & status monitoring (2010-v8)
DESCRIPTION
This is a variant of a talk that I gave at Predictive Analytics World in February 2010.TRANSCRIPT
1
Health & Status Monitoring: Two Case Studies
Robert Grossman Open Data Group
February 18, 2010
2
1. Introduction
3
Traditional Approach
• Two types of variation:– Common cause of variation (noise) occur as
normal part of manufacturing process– Special cause of variation represents a
potential problem
4
Shewhart control chart used by NIST for calibrating the standard KG.
3 s
Source: NIST
5
Shewhart / Deming Cycle
• Plan – identify opportunity or problem and make a plan.
• Do – implement the change on a small scale and collect the data.
• Check – perform a statistical analysis and check if there was an impact.
• Act – if there was an impact, broaden the scale and continuously improve your results.
6
Case Study 1. Data Center
• Thousands of servers
• Complex workloads
• Large variations are normal
• Problems make the front page
7
Case Study 2. Payments Network• Billion+ cards• 100+ million terminals• Millions of merchants• Thousands of transactions per
second• Thousands of member banks• Data highly heterogeneous
– Variations among products– Variations among cardholders– Variations among merchants– Variations among banks– Variation among payment
networks
8
The Challenge Today
• Many sources and data feeds • Data is complex and highly heterogeneous• High volume, streaming data from around
the world• Multiple parties involved, each of which
can modify the data in subtle ways
9
Health & Status Monitoring SystemsSQC HSM
What is monitored?
Single assembly line producing physical widgets
Digital system with thousands of data feeds
Type of model
Exceed 3 standard deviations
?
# of models Single model ?Visualization Control chart ?
Process Plan-do-check-act ?
10
2. The Technology
11
Observed Model
Baseline Model
CUSUM modelsGLR models
12
Build more than 104 Models: One for Each Cell in Cube of Models
Build separate model for each bank (1000+)
Build separate model for each geographical region (6 regions)
Build separate model for each different type of merchant (over 800 types of merchants)
For each distinct cube, build a distinct model
Geospatial region
Type of Transaction
15,000+ separate baselines
Modeling using Cubes of Models (MCM)
Bank
data updates1. data
collection
Operational systems, data feeds, warehouses, …
3. on-line scoring
Model Consumer
candidate alerts
features
events
PMMLmodels
Entity/Feature Database
2. off-line modeling
Data Mining Mart
learning setsData Mining
System
Rules
Dashboard engine
4. reporting
13
reports
14
Augustus• Augustus is an open source data mining platform:– Used to estimate baselines for over 15,000
separate segmented models – Used to score high volume operational data and
issue alerts for follow up investigations • Augustus is PMML compliant• Augustus scales with– Volume of data– Real time transaction streams (15,000/sec+)– Number of segmented models (10,000+)
15
Greedy Meaningful/Manageable Balancing (GMMB) Algorithm
• Fewer alerts
• Alerts more manageable
• To decrease alerts, remove breakpoint,order by number
of decreased alerts, & select one or more breakpoints to remove
• More alerts
• Alerts more meaningful
• To increase alerts, add breakpoint to split cubes,
order by number of new alerts, &
select one or more new breakpoints
One model for each cell in data cube
Breakpoint
16
3. Case Studies
17
Case Study 1
Open Cloud Testbed Monitor
18
Results• Dozens of separate statistical baselines models
developed and deployed.• Effective for discovering nodes that are hindering
effective use of OCC’s large data cloud.• Dead nodes are easy to identify and remove.• Removing just one or two “slow” nodes from a
pool of 100 nodes can improve overall performance by 15% - 20+%.
19
Dashboard
20
Case Study 2Account
Merchant
Issuing Bank
Acquiring Bank
Payments Network
21
Program Structure• Strategic objective identified early:
– “Identify and ameliorate data interoperability issues to improve the approval rate of valid transactions and the disapproval rate of invalid transactions, ...”
– Report quarterly to CIOs’ council with third-party endorsed monetary benefits summarized on an executive dash board
• Introduced data governance program early in project• Developed payment transaction monitor that produced
candidate alerts• Set up investigation process to screen alerts and
investigate those of interest• Developed reference models and appropriate standards
22
Results• ROI– 5.1x Year 1 (over 6 months)– 7.3x Year 2 (12 months)– 10.0x Year 3 (12 months)
• Over 15,500 separate statistical baselines models developed and deployed.
• Also developed appropriate rules-based models to make work of analysts more efficient.
23
4. Summary
24
Strategic Objective Dashboard
Governance
Monitor - produces candidate
alerts
events program alerts
Business Process
Modeling Process
Investigative Process
candidate alerts
Reference Model
InvestigativeProcess
25
Some Lesson Learned• Business Processes
– Importance of “C”-level executive support, dashboard reports, and a data governance program
• Modeling Processes– Critical to build as many statistical models as the data
required; used open source Augustus software for this– Architecture separated offline modeling and online scoring– Post processing with business rules to control workflow to
analysts• Investigative Processes
– It is not about the models and alerts – it is about optimizing the analysts’ workload and derived business value
– Small changes in report designs had large impact in the effectiveness of the alerts
26
SummarySQC HSM
What is monitored?
Single assembly line producing physical widgets
Digital system with thousands of data feeds
Type of model
Exceed 3 standard deviations
Change detection model
# of models Single model Cube of modelsVisualization Control chart Dashboard
Process Plan-do-check-act Plan-do-check-act
27
For More Information
• Robert Grossman– grossman.info at gmail.com– rgrossman.com (blog)
Learn about Health and status monitoring• Open Data Group• www.opendatatgroup.com
28
References• Joseph Bugajski, Chris Curry, Robert L. Grossman, David Locke, Steve
Vejcik, Detecting Changes in Large Data Sets of Payment Card Data: A Case Study, Proceedings of The Thirteenth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2007), ACM, 2007
• Joseph Bugajski and Robert L. Grossman, An Alert Management Approach to Data Quality: Lessons Learned from the Visa Data Authority Program, Proceedings of the 12th International Conference on Information Quality, (ICIQ 2007).
• Walter A. Shewhart, Statistical Method from the Viewpoint of Quality Control, Dover, 1986.
• H. Vincent Poor and Olympia Hadjiliadis, Quickest Detection, Cambridge University Press, 2009.
• Augustus is an open source system available from augustus.googlecode.com.