Sei sulla pagina 1di 31

Group 9 Earle Prithviraj DM14116 Sankuru Anil Kumar DM14203 Maddula Mahendra Avinash DM14233

EXPERTS VIEWS/DEFINITIONS OF ANALYTICS & PREDICTIVE ANALYTICS:


What is Analytics?
the application of computer technology, operations research and statistics to solve problems in business and industry. Analytics is carried out within an information system.

the application of computer technology, operations research and statistics to solve problems in business and industry. Analytics is carried out within an information system.

Tom Davenport
noted author Using analytics is like driving your car but watching traffic through the rear-view mirror, not seeing whats ahead and thereby in danger of crashing

WHAT IS PREDICTIVE ANALYTICS?


predictive models exploit patterns found in historical and transactional data to identify risks and opportunities. Models capture relationships among many factors to allow assessment of risk or potential associated with a particular set of conditions, guiding decision making for candidate transactions.
Any solution that supports the identification of meaningful patterns and correlations among variables in complex, structured and unstructured, historical, and potential future data sets for the purposes of predicting future events and assessing the attractiveness of various courses of action.

Using predictive analytics is like driving your car and watching traffic through the front windshield, anticipating traffic, making course corrections to avoid traffic jams and getting there faster and safer

WHAT IS PREDICTIVE ANALYTICS?


A set of business intelligence technologies that uncovers relationships and patterns within large volumes of data that can be used to predict behavior and events Predictive Analytics is forward looking, using past events to anticipate the future

WHAT IS PREDICTIVE ANALYTICS?


Other BI technologies are deducting in nature validating their hypotheses
Predictive Analytics is Inductive in nature. pull out meaningful relationships and patterns.

HOW ANALYTICS AND PREDICTIVE ANALYTICS COMPARE


Analytics is the understanding of existing (retrospective) data with the goal of understanding trends via comparison
Developing analytics is the first step towards deriving

predictive analytics

Predictive Analytics are more sophisticated analytics that forward thinking in nature

They used for gaining insights from mathematical and/or financial modeling by enhancing understanding, interpretation and judgment for the purpose of good decision making

HOW ANALYTICS AND PREDICTIVE ANALYTICS COMPARE


Attribute Purpose Analytics Understand the past Observe Trends Catalyst for Decision Historical and Current Lagging Indicators Raw & Compiled Predictive Analytics Gain Insights Make Decisions Take Action Future Oriented Leading Indicators Information

View Metrics type Data Used

Data Type
Benefits

Structured
Gaining an Understanding of data Productivity Improvement

Structured and Unstructured


Gaining Information & Insights Process Improvement

ANALYTICS

ANALYTICAL TECHNIQUES
Descriptive Model Prediction Model Decision Model

Find clusters of data elements with similar characteristics Focus on as many variables as possible Examples: customer segmentation based on sociodemographic characteristics, life cycle, profitability, product preferences

Find causality, relationships and patterns between explanatory variables and dependent variables Focus on specific variables Examples: next customer preference, fraud, credit worthiness, system failure

Find optimal and most certain outcome for a specific decision Focus on a specific decision Examples: critical path, network planning, scheduling, resource optimization, simulation, stochastic modeling

ANALYTICS MATURITY CYCLE

PROCESS OF PREDICTIVE ANALYTICS


Project Definition

Model Management

Exploration

Deployment

Data Preparation

Model Building

PROCESS OF PREDICTIVE ANALYTICS


Most process for creating predictive models incorporate the following steps 1. Project Definition / Business Understanding

2.

Define business objectives and desired outcomes

Exploration / Data Understanding

Analyze source data to determine appropriate data, model building approach and scope

3.

Data Preparation

Select, extract and transform data to create models

4.

Model Building

Create, test and validate models, and evaluate them

5.

Deployment

Apply model results to business decisions or processes

6.

Model Management

Manage models to improve performance, accuracy, control access , promote reuse

PROCESS OF PREDICTIVE ANALYTICS

PREDICTIVE ANALYTICS METHODS


Analysts

build models using different techniques: neural networks, decision trees, linear regression, nave Bayes, etc. Skill in creating effective analytic model is knowing which models and algorithms to use Many analytic workbenches now automatically apply multiple models to a problem to find the combination that works best. Advances make it possible for non-specialists to create fairly effective analytic models

PREDICTIVE ANALYTICS TOOLS


Widely used tools are:
Open Source Predictive Tools Commercial Tools

KNIME R
WEKA Orange

Oracle Data Mining (ODM) Minitab


SAS and SAS Enterprise Miner IBM SPSS Statistics and IBM SPSS Modeller

Rapid Miner

Oracle Data Mining (ODM)

APPLICATIONS OF PREDICTIVE ANALYTICS

Retail

Marketing

Financial Services

Online and Social advertising

Health Care and Pharmaceuticals

RETAIL ANALYTICS
Mark Down Optimization
For leading retailers, markdown optimization events provides an opportunity to maximize margin, sell-through, and inventory value while improving velocity of product assortment.

Work Force Optimization


Workforce optimization supports the business with key insights into how its workforce is performing. In the closely related area of workforce management, the emphasis is on improving operational efficiency and managing the workforce effectively while keeping overall costs at a minimum

Supply Chain Optimization


A natural offshoot of Forecasting is optimizing vendor order and store order quantitiesa trade-off between the inventoryholding and stock-out costs.

MARKETING
Audience Segmentation A brands audience can be represented and profiled at varying degree of detail. A basic overall profiling is the first step to get a very high level understanding of the audience, but it cannot answer several questions.

Market Basket Analysis Retailers want to understand which products/brands sell together (affinity) and which products/brands cannibalize each other. Applying analytics to historical POS data at he basket-level, we can track affinity and cannibalization relationships between various products/brands/categories across different countries/regions/stores. We can quantify the financial impact of these relationships, and also recommend promotional and pricing Competitive Analysis strategies specific to a product relationship. With more and more marketing budgets moving from traditional channels to social media channels, there is a clear interest among brand managers to better understand what can social do for them.

FINANCIAL SERVICES

Risk Management
Attract and grow lower-risk highly profitable customer segments . Tighten controls over credit quality, loss, pricing and exposure relative to reserve and funding requirements . Expand and strengthen the range of risk factors to refine and reduce errors in loss forecasting Examples: Application Scorecards, Fraud Management

Account and Portofolio Management


Maximize account value, minimize risks and make optimal portfolio-level decisions.

Collections:
Identify customers with temporary setbacks where refinancing can reduce losses and build loyalty . Assign treatment approaches and resource allocation relative to repayment potential . Determine which customers to negotiate settlement or refer to agencies before expending precious resources on unproductive collection attempts .

ONLINE AND SOCIAL ADVERTISING

Bid Optimization for Online Advertising


All online ad exchanges function through a dynamic auction market where marketers compete for the audience they would like to reach. Deploying optimal bidding strategies that are tailored to the end-goal of the advertiser is paramount to succeed in these markets

Customer Acquisition through Social Media


Social media provides unique ability to target potential customers based on various demographic and psychographic dimensions.

Example: A company which is a fast-growing social media start-up, could need a developed bid recommendation engine that leads to significant reductions in cost-per-click (CPC) and delivers higher post-click user engagement rates.

Example: A company acquire customers for a major credit card issuer through Facebook advertising. Predictive modeling is used to develop models that minimizes the cost per customer and acquires across a complex product and offer mix spanning major international markets.

Waiting time analysis

Length of stay analysis and optimization

Patient satisfaction analysis, reporting and improvement

Clinical performance analysis


Healthcare Applications Procedure cost analysis and optimization

Patient-level costing

Supply and demand planning

Staffing planning and optimization

TRENDS IN PREDICTIVE ANALYSIS


While

there is substantial increase in interest for Predictive Analytics in the BI community, few organizations have taken the plunge A lot of companies want to do predictive analytics, but have yet to master basic reporting Deloitte Consultings Miller Only about 1/3 of organizations say they have implemented predictive analytics in a mature fashion that uses well defined processes and measures of success that enables them to continuously evaluate and improve their modeling efforts

BENEFITS OF ANALYTICS AND PREDICTIVE ANALYTICS

Benefits of analytics: productivity gains through improved data-gathering processes results in less time required for producing reports and metrics

Benefits of predictive analytics: process improvement gains through improve revenue generation & cost structures leading to enhanced decision making

Takeaway: Both types of gains are beneficial but improvements in analytics are NOT as scalable as to the benefits in predictive analytics which are repeatable, virtuous and scalable

BENEFITS AND PITFALLS OF PREDICTIVE MODELING


Benefits of Predictive Modeling Multivariate pricing has significant advantages over traditional techniques Corrects methodological flaws Does more with limited data Provides better diagnostics to support decision-making Eliminates time-consuming adjustments Predictive modeling is being successfully applied within the insurance industry for a wide variety of applications including Pricing UW Marketing Claims

Pitfalls of Predictive Modeling Reliable Data IT Availability Shortcutting the Process Treating Predictive Modeling as a Black Box Senior Management Understanding/Buy-In

PROS AND CONS OF PREDICTIVE ANALYTICS:


Sometimes not so easy to explain; Some models are over parameterized and have too many weights and over-fit the data if you are not familiar how to avoid this danger; Could be time consuming/require computing resources; Concerns about privacy and security, regulatory issues, bandwidth for moving data to the cloud, and increased complexity

Impressive predictive power if you know how to use it; Flexible (transformations, interactions, any factors number, clusters, ) Based on good math theory;

BUSINESS PEOPLE (USERS) VS TECHNOLOGY PEOPLE (TECHIES) IN PREDICTIVE ANALYTICS:


The purpose of predictive analytics is to help organizations see relationships

between business elements so senior management may craft targeted business strategies and exploit opportunities on a timely basis with a focus on the future In order to benefit from predictive analytics, people across the organization must communicate and understand with one another but language often becomes a barrier BI professionals often think language is SQL (Structured Query Language) and business people often think language is reports, metrics and meetings IT & BI professionals need to understand the language of strategy, business models and performance while solving business not technology problems

SQL

vs

EXAMPLE CONVERSATION BETWEEN CEO AND TECHIE:


OK, what are the parameters and how do you want it rendered?

Need market segmentation report, now!

CEO/Business People

BI People

Conversations @ Work

CONTD.
Huh? What is he asking me? Just need my report!

Huh? What is he asking me?

Market Segmentation?

CEO/Business People

BI People

The Communication is very important aspect in Predictive analytics. The Perspective of Business people and Techies are very different.

PREDICTIVE ANALYTICS : MACRO AND MICRO LEVELS


Macro Level: Strategic Planning Financial Planning Focusing on Priorities Competitive Analyses Achieving Profit and Revenue Targets Developing Competitive Advantages and Differentiation
Micro Level: Improving business processes Doing more with less budget (working smarter not harder!) Allocating resources appropriately Understanding correlations and sensitivities with customer segments To ensure long term financial resources are available to run the business Developing Competitive Advantages and Differentiation

Predictive analytics can provide timely feedback to executives on their strategic initiatives without feedback course corrections may be too late Predictive analytics provide leading indicators and insight to assist in planning for answering the big question: What should we do next? next quarter, next year etc. Organizations fail to recognize and misunderstand the necessary and intangible elements of people, skills, and corporate culture and tying these elements back to their analytics, business model and strategies Caution: this is a long-term fix

THANK YOU

Potrebbero piacerti anche