Deriving Value From Big Data Visual, Predictive, GeoLocation and Event Analytics Nick Young Solutions Consultant - APJ nyoung@tibco.com
Analytics Insight to Action Value Grow Revenue Reduce Risk Analytics Insight to Action Analytics Journey Visual Analytics and Dashboards Predictive & Prescriptive Analytics GeoSpatial Analytics Event Analytics Wrap-Up / Questions Putting it all together with some examples Increase Productivity ROI
Insight to Action Batch Analysis Real-Time Analysis Process Zone Strategic Tactical Operations Execution Time Increment Quarters / Years Months / Quarters Hours / Days / Weeks Seconds/ Minutes Importance of Alerts Not Important Important Necessary Value to the Organization
Insight to Action Analytics Self-service Dashboards Predictive and Prescriptive Analytics Event Processing Measure Diagnose Predict Optimize Operationalize Automate Analytics Maturity Immediate Value to the Organization Long-Term Competitive Advantage Get business value across this Analytics Spectrum
Insight to Action Analytics Self-service Dashboards Predictive and Prescriptive Analytics Event Processing Measure Diagnose Predict Optimize Operationalize Automate Analytics Maturity Immediate Value to the Organization Long-Term Competitive Advantage Get business value across this Analytics Spectrum
Insight to Action Analytics Self-service Dashboards Predictive and Prescriptive Analytics Event Processing Measure Diagnose Predict Optimize Operationalize Automate Analytics Maturity Immediate Value to the Organization Long-Term Competitive Advantage Get business value across this Analytics Spectrum
Insight to Action Analytics Self-service Dashboards Predictive and Prescriptive Analytics Event Processing Measure Diagnose Predict Optimize Operationalize Automate Analytics Maturity Immediate Value to the Organization Long-Term Competitive Advantage Get business value across this Analytics Spectrum
Dashboards - Some Basic Charts and Plots Bar chart Line chart Pie chart Scatter Plot Bar chart Combo chart Parallel Coord Plot 3d Scatter Plot
Dashboards - Some Basic Charts and Plots Graphical Table Heat Map Box Plot Tree Map Table Cross Table
Visual Analytics - Interactive Brush-Linked
Visual Analytics - Interactive Brush-Linked Polar 3D rotate Surface Contour Network Funnel
Extending the Palette JS, D3 Sankey Venn Dials Gantt Donut Chord
Dashboards and Themes Branding
Dashboards and Themes KPIs
Dashboards and Themes Color
Dashboards and Themes Color
Map Charts
Map Layers Marker Layer Feature Layer Map Layer WMS Layer Image Layer
Map Elements Marker Layer Color Shape Size Feature Layer Color Relative amounts Size Marker or Feature Layer Tooltip Labels
Non Geographic Data Interactive Charts
Non Geographic Data Interactive Charts Airplane Seating Chart Football Field Seating Chart
Pixel-Perfect Embedded Reports
From Dashboards to Value Analytics Self-service Dashboards Predictive and Prescriptive Analytics Event Processing Measure Diagnose Predict Optimize Operationalize Automate Analytics Maturity Immediate Value to the Organization Long-Term Competitive Advantage Get business value across this Analytics Spectrum
Big Data Analytics
Data a Insight a Action Business Case Prepare Data Explore Data Guided -------- Deploy -------- In-Line Improve Value Theses Data Access & Prep Exploratory Data Analysis Features Visual Dashboard Model & Predict Deploy Model Test & Learn Increase Productivity Grow Revenue Reduce Risk ROI Purchase Campaign Loyalty Churn Channel At Rest In Motion Visualize Aggregate Shape Merge Map Filter Big Data Propensity Affinity Segment Promotion Pricing Ensemble Forest Regression Additive Models Fast Data Dashboard Updates Challenger Models
Exploratory Data Analysis (EDA) & Features Features: Telco Churn April 21 Customers Business Problem: who is likely to leave? Insight: people who spend time talking to people who have already left the plan are more likely to leave!
Exploratory Data Analysis (EDA) & Features Features: Telco Churn May 2 Deactivations Business Problem: who is likely to leave? Insight: people who spend time talking to people who have already left the plan are more likely to leave!
Exploratory Data Analysis (EDA) & Features Features: Telco Churn June 4 Deactivations Business Problem: who is likely to leave? Insight: people who spend time talking to people who have already left the plan are more likely to leave!
Exploratory Data Analysis (EDA) & Features Features: Telco Churn July 7 Deactivations Business Problem: who is likely to leave? Insight: people who spend time talking to people who have already left the plan are more likely to leave!
Key Feature Features: Telco Churn Feature: % incoming calls in last <90> days from customers who cancelled in last <30> days Those with >10% = churn chatters Feature not in any database! Voice a little better than text Call count a little better than call length Inbound closer than outbound
Features Telco Churn Calls to Call Center Time Left in Plan (days) Tenure (days) Time Since Last Call (days)
Social Amplification of Features / Not Long to Act! Time Left in Plan (days) Red = churn chatters Comparing Churn Rates Over Time Pct Churning in Each 30-day period 0.00 0.05 0.10 0.15 Feb Mar Apr May Churn chatters gone in 90 days All EM/EM+ Subs Top 5% Per Base Model Top Churn Chatters
R the Lingua Franca of Data Science
Contextual + Simple Predictive Analytics Contextual Analytics - Right click - Forecasting Contextual Analytics - Menu - Machine Learning
Extensible Predictive Analytics Expressions Lines by TERR( contourlines(x,y) ) Interactive Advanced Analytics - Expression on the fly R code - Points: kmeans(), - Lines: contourlines(), spline(), - Expression Functions sharing Color by TERR( kmeans(x1,x2) )
Trade Area Analysis Customers within drive/walk from location (store) Store open/close effects Competitor response Expected sales*products from demographics
Analysis Workflows Data Functions Variables driving segments - Random Forest Revenue by product - Color by segment Interactive Analytics with R - Data Function - Robust Cluster Analysis - Any Analysis in R / CRAN
Customer & Marketing Analytics Analytics Data Segmentation Propensity Affinity Cross-Sell, Up-Sell Churn Test & Learn Customer Acquisition Customer Lifecycle Relationship Growth Customer Attributes Purchase History Social, sentiment Location Satisfaction (NPS) Customer Retention
Segmentation: @CIBC
Propensity: Respond to Credit Card Offer Response Data Acceptance of credit card offer Feature Variables Bank transactions: ATM, checking; deposit, withdrawal, balance, Demographics: sex, age, profession, nationality Customers Identified Likely to respond to credit card offer
Cross-Sell: @Citibank Citibank saw time-to-market for new product offerings cut by 80% (from 8 weeks to 8 days) after implementing cross-sell programs.
From Dashboards to Value Analytics Self-service Dashboards Predictive and Prescriptive Analytics Event Processing Measure Diagnose Predict Optimize Operationalize Automate Analytics Maturity Immediate Value to the Organization Long-Term Competitive Advantage Get business value across this Analytics Spectrum
Real Time - Data Customer Analytics Customer Analytics Offer Analysis Customer Experience Campaign Effectiveness
Customer Overview & Profiles 48
Segments and Offers 49
Customer Profiles and Targets 50
Customer Profiles: Sarah to get Free Gift 51
Login Event Capture All Login events tracked and monitored.
Intent to Cancel Intervention Search events evaluated against Analytics Rules and Models.
Offer to Customer trying to Cancel (Sarah) Customer is given an offer based on their profile and actions.
Notify Call Center and Sales Team Web site data sent to Call Center and Sales Team.
Campaign Effectiveness 56
Real Time - Data Customer Analytics Customer Analytics Offer Analysis Customer Experience Campaign Effectiveness
Real Time Data Analytics Interventions Give customer service Make offer to customer Proactively maintain machines Restock inventory Optimize pricing Check for fraud Reroute transport
Wrap Up Thank you! Nick Young Solutions Consultant, APJ TIBCO nyoung@tibco.com