Business Intelligence: Using Data for More Than Analytics 2015... · ANNUAL EDUCATIONAL CONFERENCE...
Transcript of Business Intelligence: Using Data for More Than Analytics 2015... · ANNUAL EDUCATIONAL CONFERENCE...
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Business Intelligence: Using Data for More Than Analytics Session 672
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Session Overview
Business Intelligence: Using Data for More Than Analytics What is Business Intelligence? Business Intelligence Solution Data Cleansing & Data Validation Data Consolidation Using Data for More Than Analytics Question & Answers
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Introductions
Your Speakers Robert Clark
• Vice President of Development, 4Sight Business Intelligence Matt Carter
• Data Warehouse Analyst, Tower Hill Insurance Group Michael Paparatto
• Reporting and Business Intelligence Developer, Anchor General Insurance Group
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
What is Business Intelligence?
Definition:
“Business intelligence (BI) is the set of techniques and tools for the transformation of raw data into meaningful and useful information for business analysis purposes… The goal of BI is to allow for the easy interpretation of these large volumes of data. Identifying new opportunities and implementing an effective strategy based on insights can provide businesses with a competitive market advantage and long-term stability.” - Wikipedia
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Business Intelligence Solution
Three Main Components Extract Transform and Load (ETL) – BI loading process Data Store – Data storage for analysis BI Tools – End-User analysis
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Data Warehouse
Components of a BI Solution
Load
Transform
Extract
Business Intelligence Application(s)
Reports
Dashboards
Ad-hoc Analysis
Source Systems
Extract, Transform and Load
Data Store BI Tools
IOIOIOIOI IOIOIOIOI
Policy Admin
Claims Admin
Accounting
IOIOIOI
IOIOIOI
IOIOIOI
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Components of a BI Solution
Load
Transform
Extract
Source Systems
Extract, Transform and Load
IOIOIOIOI
Policy Admin
Claims Admin
Accounting
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Business Intelligence Solution
Extract, Transform and Load (ETL) Extract
• Pull data in raw format from the source system • Prepare it for transforming
Transformation • Data Cleansing (more on this shortly) • Code Translations • Calculations • Data Validation (more on this shortly)
Load • Loading the transformed/converted data into the target system or
database
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Data Warehouse
Components of a BI Solution
Load
Transform
Extract
Source Systems
Extract, Transform and Load
Data Store
IOIOIOIOI IOIOIOIOI
Policy Admin
Claims Admin
Accounting
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Business Intelligence Solution
Data Store Data Warehouse
• Star Schemas – comprising of fact tables (measures and values) connected to dimensions (values used to slice and dice the data)
• Often Contain Star Schemas with Different Levels of Aggregation (coverage, policy, claim, month, quarter, year, etc.)
Big Data • Large Datasets – datasets that are too large and complex that traditional
databases are inadequate to process the data
• Term Often Used Interchangeably with Predictive Analytics
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Business Intelligence Solution
Data Store (continued)
Cube (Online Analytical Processing – OLAP)
• MOLAP (Multidimensional Online Analytical Processing) – Data stored in N dimensional cube and calculations are generated ahead of time
• ROLAP (Relational Online Analytical Processing) – Similar to MOLAP but uses relational database to retrieve values
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Data Warehouse
Components of a BI Solution
Load
Transform
Extract
Business Intelligence Application(s)
Reports
Dashboards
Ad-hoc Analysis
Source Systems
Extract, Transform and Load
Data Store BI Tools
IOIOIOIOI IOIOIOIOI
Policy Admin
Claims Admin
Accounting
IOIOIOI
IOIOIOI
IOIOIOI
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Business Intelligence Solution
Business Intelligence Tools Data Retrieval & Basic Analysis
• Reporting and Querying Software – basic report design tools and database query clients
• Spreadsheets Dashboards – graphical snapshot of data showing historical trends and analysis
Advanced Analysis • OLAP (Online Analytical Processing) – slicing and dicing, drill down,
drill up, aggregations, etc. • Data Mining – discovering patterns in large sets of data (groups,
anomalies and associations)
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Data Warehouse
Components of a BI Solution
Load
Transform
Extract
Business Intelligence Application(s)
Reports
Dashboards
Ad-hoc Analysis
Source Systems
Extract, Transform and Load
Data Store BI Tools
IOIOIOIOI IOIOIOIOI
Policy Admin
Claims Admin
Accounting
IOIOIOI
IOIOIOI
IOIOIOI
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Data Cleansing & Data Validation
Data Cleansing is the process of identifying and correcting inaccurate or invalid data
Data Validation
is the process of using rules or constraints to check the validity or correctness of the data
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Data Cleansing
Data Cleansing Steps: Identification of Invalid Data Correcting Data or Removing It
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Data Cleansing: Identifying Invalid Data
Identifying Invalid Data Using Tools or Queries to Identify
• Field Data Types • Uniqueness Constraints (e.g. duplicate records or field values)
• Formats and Patterns (e.g. mm/dd/yyyy MM:hh:ss.SSSS)
• Ranges (min / max)
• Accuracy (e.g. address verification)
• Set Membership (e.g. Male/Female, AR/AZ/OH/NJ/TX, etc.)
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Data Cleansing: Correcting/Removing Data
Correcting or Removing Data Data is Corrected if Possible
• Invalid Values Researched to Identify Correct Values • Values Updated to “Correct” Values in Batch (using queries or tools)
• Not Possible to Bulk Update? • Correct data in the source system • Default to a value • Remove Record
Removing Data • Filter Out When Reading Data • Delete the Record
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Data Validation
Data Validation Steps Define Validation Rules Execute Rules Against Data (as loaded or in batch during ETL)
Review Invalid Data & Decide Course of Action
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Data Validation: Define Rules
Validation Rules Rules Are Written In ETL Tool, Using Queries or a Rules
Engines Example Rules Include
• All Coverage Effective/Expiration Dates Must Be Within the Policy Term
• Claim Must Be Associated with a Valid Coverage (especially if separate systems for policies and claims)
• Coverage In-Force at Time of Loss • Appropriate Coverage for Type of Loss • Sum of Loss Reserves is a Positive Number • Endorsement Prorated Premium Calculated Correctly
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Data Validation: Execute Rules
Rule Execution Rules Are Executed By
• ETL Process During Transformation Phase • Manually Through Queries or Tools • Scheduled
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Data Validation: Review & Decide Course of Action
Review the Results and Decide Course of Action to Correct Needs To Be Corrected In Source System
• Bulk Update of Data • Add Edits Source System to Prevent Issue
Needs to be Corrected Outside Source System • Same Process as Data Consolidation: Correcting or Removing Data
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Importance of Data Cleansing & Data Validation
Key Business Decisions Are Being Made on the Data Simply Just Pushing Data From Source Systems Into a BI
Solution is Not Enough • Operational Systems Are Not Perfect • Often Legacy Data Has Been Ported From Older Systems • Data is Coming From Disconnected Systems (policy, claim, TPA)
• ETL Itself May Have Issues Garbage In, Garbage Out (GIGO)
“… in the field of computer science or information and communications technology [GIGO] refers to the fact that computers, since they operate by logical processes, will unquestioningly process unintended, even nonsensical, input data (‘garbage in’) and produce undesired, often nonsensical, output (‘garbage out’).” - Wikipedia
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Data Consolidation
What is Data Consolidation? Collecting and Integrating Data From Disparate Systems
Into a Single Data Store.
Data Warehouse Policy Admin
(Personal Lines)
Policy Admin (Commercial Lines)
TPA Claims Extract
Legacy Policy Admin
Accounting
Billing System
10101010101010110101 10101010101010110101
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Data Consolidation
What Are the Challenges? Missing Data Between Systems Different Systems, Different Codes Consolidating and Referencing Entities
• Agents / Producers • Third Parties (Loss Payee, Other Interests) • Insureds • Claimants • Underwriters • Adjusters
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Data Consolidation
What are the Benefits? Centralized Reporting
• All Reports Come From One System • No Longer Need Spreadsheets to Merge Data
Single Version of the “Truth” • Calculations are Consistent Across the Organization
• Earned / Unearned Premium • Incurred Formulas • Loss Ratios • Many More
• Non-Redundant Data (one source) • Definitions and Included/Excluded Data are Consistent
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Data Consolidation
What are the Benefits (continued)? Consistent Granularity of Data
• Earned Premium (Calculated at Policy Level vs. Coverage Level)
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Using Data for More Than Analytics
You have cleaned up and consolidated all this data. Now what can you do with it?
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Using Data for More Than Analytics
Business Intelligence Standardized Reporting
Dashboards
Ad-hoc Analysis
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Using Data for More Than Analytics
Governmental Reporting Statistical Reporting
• Extracting Data for ISO, NISS, NAIC, NCCI, etc. • Third Party Submissions
State Data Calls • Extract Data in State Specific Formats
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Using Data for More Than Analytics
System Conversions Load Legacy Data Into Data Warehouse Load New System’s Data Into Data Warehouse on a Regular Schedule Provide Seamless Reporting
• No Need to Consolidate Reports • Single Version of the “Truth” • Ease Transition when Renewing into New System
Provides Coverage Verification • Coverages Can be Verified Even if Policy Is Not In New System
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Using Data for More Than Analytics
System Conversions (continued) Conversion of Data
• Data Has Been Cleansed and Validated • Convert Data Over to New System • Companies Offer Web Services To Extract Data
Conversion All At Once • Port All Data Over to New System
Conversion On Renewal • Port All Inforce Policies to New System • Then Port Policies Coming Up for Renewal
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Using Data for More Than Analytics
System Conversions (continued) Conversion of Open Claims
• Port Only Open Claims • Port Any Claim That Reopens When Needed
Conversion of Insurance Fund Assumptions (e.g. Citizens) • Loading Insurance Fund Assumptions Allows for Analysis • Conducive to Converting from DW to Operational Systems
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Using Data for More Than Analytics
Self Service and Web Portal Customer Portals
• View Policy and Coverages Agent Portals
• Ability to Query Upcoming Renewals • View Profitability • Of Course Check Commissions
Third Party Portals (e.g. TPAs, Reinsurers, Auditors, etc.)
• View and Query Data • Controlled Analysis of Data
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Slide Title Goes Here in 28pt Arial Bold Flush Left
Predictive Analytics Machine Learning – algorithms that can learn and make predictions on the data (Open Source R)
Predictive Modeling – using statistics to forecast or predict outcomes
Data Mining – discovering patterns in large sets of data (groups, anomalies and associations) Examples of Predictive Analytics • Fraud Detection • Risk Management • Underwriting • Cross Selling
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Using Data for More Than Analytics
Other Uses CLUE & APLUS Claims Contributions and Submissions DMV Reporting OFAC Validation Medicare Regulations (e.g. 111) Reinsurance Reporting
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Conclusion
Using Data for More Than Analytics Business Intelligence Solution
• Extract, Transform & Load • Data Store • Business Intelligence Tools
Importance of Data Cleansing & Data Validation • Key Business Decisions Are Being Made on the Data • Simply Just Pushing Data From Source Systems Into a BI Solution is
Not Enough. It Needs to be Validated and Cleansed. Consolidation of Data
• Single Version of the “Truth” • Seamless Reporting
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Conclusion
Using Data for More Than Analytics Uses For the Cleaned & Validated Data
• Governmental Reporting • Statistical Reporting • State Data Calls
• System Conversions • Conversions of Data all at Once or On Renewal • Conversion of Open Claims and Reopen When Needed
• Self Service Web Portals • Customers • Agents and Third Parties
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Conclusion
Using Data for More Than Analytics (Continued) Uses For the Cleaned & Validated Data
• Predictive Analytics • Machine Learning • Predictive Modeling • Data Mining • Fraud Detection, Cross-Selling, Risk Management, etc.
• Many Other Uses • CLUE & APLUS Claims Contributions and Submissions • DMV Reporting • OFAC Validation • Medicare Regulations (e.g. 111) • Reinsurance Reporting
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Question & Answers
Questions?
IASA 87TH ANNUAL EDUCATIONAL CONFERENCE & BUSINESS SHOW
Please Complete the Session Evaluation Form on the Conference App