PPT Presentation

Module Overview | Careers in Analytics In this module, we will evaluate the various quantitative data collection and analysis methods in

standard industry practice. These methods are what will be used throughout this program, so

you should become familiar with the terminology.

The second part of this module presents a variety of career paths for data analysts and an

overview of how several industries are currently using data analytics. Pay special attention to

the intersection of skills necessary for a data analyst to possess, and think of the steps you can

take to gain or improve on these in your own skill set. This may give you an idea of the career

path and industry you would like to pursue, or enhance your understanding of a career path and

industry you have already chosen.

Industry Practice

Learning Objectives

Explain the technical elements and steps associated with analytics practices and processes

Explore industry practice of data analytics

Typical Quantitative Techniques Used in Advanced Analytics

Several quantitative techniques apply to analytics projects, including:

Type Description

Simulation Randomized repetitions of a set of discrete events in order to

model real-world systems and phenomena (e.g., queues)

Optimization Algorithm selects the best possible outcome, subject to

satisfying constraints

Matrix Algebra Calculations involving matrices solve multidimensional

problems

Fitting Functions to Data Also called “curve fitting,” using numerical methods to

interpolate data

Survival Analysis Originally used by life scientists, but adopted by marketers and

actuaries

Time Series When data are “auto-correlated,” such as time-dependent data

(also called “Box-Jenkins”)

Predictive Analytics and Machine Learning

Classical Statistics

Descriptive: calculates metrics to characterize the distribution of values of data (mean, standard deviation, range, etc.)

Predictive: estimates parameters using historical data and making predictions of future outcomes (multivariate regression,

generalized linear regression, etc.)

Learning

Unsupervised learning: characterizes the data to establish classes without using explicit metrics, e.g., k-means clustering

Supervised learning: Classify and describe the data with pre- defined ‘labels,’ e.g., decision trees

Bayesian Used to augment classical analysis when there is prior

knowledge about how the data was generated

Typical Challenges and Pitfalls in an Analytics Project

1. Poorly defined problem

• Unclear goal of problem-solving

• Scope is unclear, e.g., how many SKUs to analyze

• Mixed objectives, e.g., economic analysis of a product category promotion for retailer versus

CPG mixed

2. Limited IT resources

• Cloud data can’t be acquired off-line within a reasonable time

• Can’t run the complete model due to computation limitation

• Too slow to generate results in real time

• Can’t share the data and results with network limitation

3. Less-best approach

• Selected less effective modeling method

• Incremental accuracy doesn’t offset the extra complexity

• Inadequate or incorrect performance monitoring criteria

4. Incomplete or incorrect data

• Primary dataset unavailable

• Complementary data unavailable, e.g., missing competitor pricing data

• Coarse data or aggregated data

• Very sparse data with missing values

5. Insufficient communication

• Insufficient data dashboard to communicate the analysis result

• Lack of soft skills to sell the results and insights

• Long feedback cycle to make the results less relevant

• Isolated org structure to stifle collaboration

Learn by Doing

Of the typical quantitative techniques used in advanced analytics, the types used in machine

learning and predictive analytics include: (Select all that apply.)

 Predictive statistics, which estimate parameters using historical data and making

predictions of future outcomes (multivariate regression, generalized linear regression,

etc.)

 Optimization, in which an algorithm selects the best possible outcome, subject to

satisfying constraints

 Supervised learning, which classifies and describes the data with pre-defined ‘labels,’

e.g., decision trees

 Descriptive statistics, which calculate metrics to characterize the distribution of values of

data (mean, standard deviation, range, etc.)

Answers:

Supervised learning, which classifies and describes the data with pre-defined ‘labels,’ e.g., decision trees; Descriptive statistics, which calculate metrics to characterize the distribution of values of data (mean, standard deviation, range, etc.)

Careers in Analytics

Learning Objectives

Explain the technical elements and steps associated with analytics practices and processes

Connect the context of analytics to marketing, risk, financial, etc., within a company and industry

Many skill sets come to play during the course of a data analysis project workflow. These

include hacking skills, math and statistics knowledge, and substantive expertise. Whereas

traditional research relies primarily on math/statistics and domain expertise, modern data

science typically draws from all three sets. The hacking skills reflect themselves in the

understanding of available tools and technologies.

Full Suite of Data Scientist Skill Set

Technical area Technology Academic discipline Domain knowledge

Data Mining

Predictive Modeling

Machine Learning

NLP

Text Analytics

Data storage and

processing

Computing

environment

Computer

programming

Visualization

BI / reporting

Probability Theory

Statistics

Computer Science

Operations

Research

Economics

Healthcare

Retail marketing

Financial services

Manufacturing

Telecommunication

Rapid Growth Projected in Big Data Market

Wikibon projects the Big Data market will top $84B in 2026, attaining a 17% Compound Annual

Growth Rate (CAGR) for the forecast period 2011 to 2026.

Line chart describing Wikibon Big Data Market Forecast 2011-2026 ($US B)

The Big Data market reached $27.36B in 2014, up from $19.6B in 2013. These and other

insights are from Wikibon’s excellent research of Big Data market adoption and growth. The

graphic above provides an overview of their Big Data Market Forecast.

Source: Executive Summary: Big Data Vendor Revenue and Market Forecast, 2011-2026.

Note–Amazon’s annual revenue is about $100B

Significant Gap in Data Analytics Demand and Talent

The demand for deep analytical talent in the United States could be 50 to 60 percent greater

than its projected supply by 2018.

The demand and supply are projected to be noticeably mismatched in the 2018 forecast. This

would be a striking change from 2008 when 150,000 slots were filled by ‘Data Analytics’

graduates with an exceeded supply of 30,000. In 2018, the projected demand (est. over

440,000) is expected to exceed the projected supply (est. 300,000) by 140,000.

Learn by Doing

The main skill set for a Data Scientist consists of: (Check all that apply.)

 Science Knowledge (e.g., Chemistry, Genetics, etc.)

 Technical Area (e.g., Data Mining, Machine Learning, etc.)

 Academic Discipline (e.g., Probability Theory, Statistics, etc.)

 Domain Knowledge (e.g., Healthcare, Financial Services, etc.)

 Technology (e.g., Computing Environment, Data Storage and Processing, etc.)

Correct Answers:

Technical Area; Technology; Domain Knowledge

Module Summary | Careers in Analytics

In this module, we presented a terminology overview of quantitative data collection and analysis

methods that you will become more familiar with as you progress through this program. These

methods are used in common practice throughout most industries and are enhancing the data

collection and analysis trends of today.

We also discussed a variety of career paths for data scientists and an overview of how several

industries are currently using data analytics. We learned about the growing demand for data

analytics talent and the talent gap that exists in the field. To be a part of that talent demand,

there is a vast skill set that data analysts need to possess, and throughout this course, we will

examine each of them in more detail, while giving you the tools you need to gain or enhance

these skills for your future career in data analytics.