Delta Model and The Three V's

Delta Model

• Data should be clean, unique, and accessible
• Enterprise-wide focus involves key data system and analytics resources available across the firm, not isolated to teams
• Leaders at all levels should promote a data and analytic driven culture
• Targets should be aimed at key business areas that could benefit from the data analysis approach
• Analysis should be talented to execute the strategy in a complex and dynamic business environment
• Big data and smart analytics improve decision making and ensure a company stays ahead of competitors
• Prescriptive analytics is a technique that uses data to recommend actions which improve business outcomes

The Three V's

• Volume: Involves various data sources such as transactions, IoT, social media, videos, and audios
  • Handled with cloud, Hadoop, and data lakes
• Velocity: Involves RFID tags, smart meters, and sensors producing data at unprecedented speeds
• Variety: Includes text, emails, documents, videos, and audios

Extracts, Transform, and Load (ETL)

• Core principles of ETL include identifying high-quality data free from integrity and consistency errors from various sources
  • Sources include data warehouses, data marts, data lakes, transaction databases, external websites, or mobile data
• Transactions in ETL target a repository creating a "single source of truth" while applying data governance guidelines
• ETL provides easy access and the capability to use the data repository in support to analytics related work

Meta Data

• Business Metadata: Contains data ownership information, business definitions, and charging policies
• Technical Metadata: Includes database system names, table and column names, sizes, data types, and allowed values, plus structural information
• Technical Metadata also includes primary and foreign key attributes and indices
• Operational Metadata: Includes currency and data lineage which refers to whether data is active, archived, or purged, plus the history of data migration and transformation

Data Analytics During COVID-19 Pandemic at Parkland Hospital

• Real-Time Dashboards: Near-real-time dashboards monitor bed occupancy, ventilator usage, and patient statistics, enabling rapid response
• Predictive Analytics: Anticipated patient admission trends and resource needs, facilitating proactive planning and resource allocation
• Enhanced Decision-Making: Access to real-time data improved decision-making, allowing swift adjustments to emerging trends and challenges

How Brands Use Big Data Analytics

• Boosting Customer Acquisition/Retention: Big data analytics helps understand customer behavior and preferences to tailor products and services
• Driving Product Development: Analyzing market trends and customer feedback to identify opportunities for new or improved products
• Personalizing Customer Experience: Big data enables personalized marketing campaigns and product recommendations
• Optimizing Operations: Analyzing operational data allows companies to streamline processes, reduce costs, and improve efficiency
• Enhancing Risk Management: Big data analytics helps identify potential risks by analyzing patterns and anomalies

Structured vs Unstructured Data in Educational Institutions

• Unstructured data often includes lecture recordings, emails, social media interactions, research papers, surveys, images, and photographs
• This unstructured data is often used to enhance teaching methods and to monitor student engagement
• Unstructured data can be used to conduct research and improve campus services by analysing survey responses
• Unstructured data can create business value through personalized learning experiences, predictive analytics for student success and enhanced alumni engagement
• Institutions can tailor educational content to individual student learning styles to improve academic performance and student satisfaction, through data analysis and feedback
• Examining patterns in emails and social media interactions can identify students at risk of challenges, allowing interventions and support
• Leveraging unstructured data from social media and correspondence can strengthen relationships with alumni, leading to fundraising and mentorship opportunities

Did You Know 3.0

• It emphasizes continuous adaptation to technological advancements and developing digital literacy skills
• Lifelong learning is important in an ever-evolving digital landscape
• Businesses are recognizing the accelerating pace of technological change as impacting consumer behavior and market dynamics
• Incorporating data analytics, innovation, and maintaining agility helps businesses stay competitive
• Understanding global connectivity and information flow ensures more informed and timely decisions
• Integrating technology into curricula prepares students for a digital future

Analyzing Customer Experience

• Descriptive analytics summarize past data to identify trends and key factors and impact customer satisfaction
• Key performance indicators (KPIs) are identified based on existing customer data
• Predictive models are used to classify the customer experience (positive or negative) based on past data and details of transactions
• Data driven stimulations and recommendations help decide a course of action
• Controlled testing and stimulations can compare customer engagement before and after implementation and analyzation of trends

Competing on Analytics

• Data analytics provides competitive advantage
• Analytics driven companies lead decision making
• Companies like Amazon, Capital One and Harrah’s now use analytics to optimize decisions

Competing on Analytics - Why

• Businesses today have vast amounts of data and analytics due to high data availability
• Industries offer similar products, making analytics driven process optimization a key differentiator
• Analytics based decisions help improve pricing, inventory management, customer experience and employee performance

Competing on Analytics - Characteristics

• Data driven decision making occurs at every level
• Data modeling and optimization is used to identify trends
• CEO's champion analytics in the work place such as Jeff Bezos from Amazon
• Empolyees rely of facts rather than intuition in a data driven culture

Competing on Analytics - Real world examples

• Capital One runs over 30,000 experiments annually
• Progressive Insurance uses predictive analytics for risk categories
• Mariott uses analytics for dynamic pricing and revenue management, improving revenue capture from 83% to 91%
• Amazon hires top analytics experts to optimize supply chain management

Competing on Analytics - Key areas

• Predictive models optimize pricing and promotions in marketing
• Supply chain management monitors real time data and allows for re-routing of shipments
• Operations optimization helps companies use multi-year data sets to optimise spending on advertising

Competing on Analytics - Approach

• Centralized data management and data insights are key to preventing discrepancies in business intelligence
• Ensuring all departments benefit with cross-functional analytics teams
• Embedded analytics into decision making across multiple business units is key

Competing on Analytics - Culture

• Transforming analytics to a core strategy across the company is a must
• There has been a move from tenure based rewards to performance based incentives, which is a change in culture

Competing on Analytics - Talent and Technology

• The biggest key is hiring and analyzing analytical talent
• Seeking analysts with consulting and financial backgrounds
• Advanced technology investments is key such as AI driven analytics and data warehouses

Competing on Analytics - Journey

• It can take years to integrate analytics
• Companies must accumulate years of historical data
• Becoming a competitive advantage is key to embracing analytics early on

Competing on Analytics - Summary

• Improving data analytics is key for leading businesses
• Culture shift and strong leadership is required to compete

Definition of Data Quality

• It is "fitness for use", depending on the needs of data consumers.
• It is subjective, relying on what one person needs

Data Quality - Characteristics

• Accuracy: Being Correct
• Consistency: No conflict in data
• Entirety: Measure of events
• Breadth: Amount of information
• Completeness: Meaure of missing data
• Uniqueness: Measure of unnecessary data replication
• Interpretability: Application of semantic standards
• Timeliness: How current
• Precison: Exactness of data
• Depth: Amount of data retained
• Integrity: Validity of data

Data Quality - Why the Struggle?

• Managing challenges in data leads to struggling within business
• Data inaccuracy is a common issue
• Optimisation on data is not prioritised
• Top reasons for data quality issues: human error, lack of communication
• Inadequate strategy for data

Cost of Poor Data Quality

• Trillions of dollars are lossed annually
• Poor data quality results in 15-25% of annual revenue loss

Data Quality - Techniques and tools

• Data profiling, monitoring, standardisation and data cleansing are all common techniques and tools

Data Quality - Data cleansing

• Data quality means understanding the need for certain information and standarding across the board

Data Quality - Bad data costs

• Bad data impacts costs hugely such as employee time and inefficient processes
• Bad data can also lead to a lack of comms internally, untrustworthy processes, and issues with system integration

Data Quality

• Assessing data quality is key with employee training and data goverance in place
• Data should be protected like an asset to make decisions based off information

Data Quality by Jack Vaughan

• It comes down to accuracy, consistency, completeness, reliability and timelines, with all employees taking resposnbiliy
• Poor data can lead to negative impacts on AI and ML There are also differnces with data integrity
• Improving data quality comes down to fixing any issues
• It can be measured and implemented across data for assessability