to

Agenda

• Introduction to HR Analytics • HR Transformation with Analytics

Overview Descriptive, Correlational & Predictive HR
Perspectives of HRM Perspectives of HRM

• Current Scenario of HR Analytics

• Analytics vs Statistics Aligning HR analytics with business goals
Introduction to Analytics Python for HR Analytics
Perspectives of HRM
Data-driven HR

• Understanding Analytics wave

Evolution & Measurement of HR
Perspectives of HRM
Data-driven HR
 Introduction
• Human resource management (HRM) is the most important function in organizations in terms
of competitive advantages, strategic relevance and sustainability.

• HRM as a function has developed primarily to promote the concept of humanization of work.

• Organizations, across the globe, are now facing the challenges of technological change, rising
competition, the rise of consumerism, structural change in employment relationships and
various other macro-economic and social changes.
• All these changes redefined the process of managing human resources, making employees as
partners of the organizations and their management as a business-aligned function.
 Different Perspectives of HRM
• HRM is a process of shaping employment relationships to achieve individual,
organizational and societal goals. This requires managers to relate their
decisions with the prevailing policies and practices.

• HRM concepts cover all HRM functions, like; human resource planning (HRP),
recruitment, employee relations, compensation management, training and
development, performance management, motivation, organizational change
and development, team development, work-life balancing, employee
engagement, talent management and so on.
• Other perspectives of HRM are descriptive and conceptual.

• The descriptive perspective of HRM emphasizes on fact-based HRM practices.

• Conceptual perspective helps us to relate facts to each other.

• HRM perspective can also be examined from strategy, process, systems, human capital
and control perspectives. At every stage, we find the need for HR analytics for
improved HR decision-making process.
 Introduction
• Globally, organizations are shifting their focus on evidence-based management practices.

• Evidence-based management practices facilitate conscientious decision-making, with

effective use of available evidences (data support) so that the decision’s outcome can
benefit the organization and the stakeholders.

• In HR we make investments in building capabilities of the organizations and employees.

We invest in attracting and retaining talent to sustain in the competition. With HR
analytics such decisions in organizations can be more and more evidence-based, leading
to more realistic cross-functional approach.

• Analytics based HR decisions, being evidence-based, are more value adding, futuristic
and sustainable.
 Business Analytics and HR Analytics
• Analytics are defined as scientific data manipulation.

• Business analytics (BA) is a scientific data manipulation for better business decisions.

• BA literally indicates application of mathematical and statistical techniques.

• Use of BA in HR decision-making process is of recent origin.

• With big-data enabled BA organizations’ decisions are more predictive.

• BA tools are more subject and function specific, of which HR analytics is one of such example.
 Different Phases of Development of
Data-driven HR Decision
• Different phases of development of data-driven HR decision-making can be understood from the
following figure:

• Use of metrics is identified as the phase of descriptive analytics; use of correlation (to study the
relation between different variables) is identified as correlation analytics, while the HR analytics
in its current form is known as prescriptive/predictive analytics.
Understanding the Analytics wave
Transforming HR with Analytics
HR Analytics and HR value propositions
• Important HR value propositions that can be objectively measured by HR analytics are:
1. Effective talent management, in terms of talent attraction, recruitment, development
and retention, for enhanced organizational performance.
2. Effective human resource planning.
3. Effective management of employee attrition, duly identifying the reasons, and initiating
the corrective actions.
4. Effective career planning and development.
5. Effective management of recruitment function.
6. Effective performance management.
7. Effective management of compensation and benefits programme.
8. Effectively positioning HR as an important enabler for driving business performance.
 Descriptive HR Decision-making
• Descriptive HR decision-making process makes use of metrics or human resource
information systems (HRIS) to get insights on decisional issues and then take decisions.

• For example, high employee attrition rate is an indicator of poor employee engagement
programme in the organization, may be for less competitive compensation and benefits,
less opportunities for employees’ growth and development and so on.

• HR managers once getting a higher rate of attrition, as per the conventional practices of
descriptive decision-making, may check above listed areas and correct the situation,
wherever required.

• Many organizations are now strengthening their HRIS with HR analytics, i.e., the process
of shifting from descriptive and correlational decisions to analytics based decision-
making process.
Correlational HR Decision-making
• Correlational HR decisions help us to assess the determinants of key variables. It
refers to study of the strong or accurate relationship, to assess the goodness of fit.

• Based on correlation HR managers can understand the possible effect of their

decisions and accordingly can calibrate their decision, to minimize the risk of adverse
consequences, if any.

• Correlation also helps in making changes in the decision for eliciting better results.
 Predictive HR Decision-making
• Predictive HR decision-making process rests on big-data analysis.

• Two commonly used predictive HR decision-making tools are causation and regression analysis.

• Through regression analysis, we estimate the trend pattern and accordingly make our decision
more robust. Multiple regression analysis can study the relationship between or among
variables, thus can influence the HR decision-making process with more precision and level of
accuracy.

• Causation analysis is different from correlation analysis. Through correlation we can understand
the relationships between two variables, but that does not speak that one causes the other.

• For example, less pay may correlate with poor productivity, but we cannot say poor productivity
is caused by less pay.

• For this we require the hypothesis framing and testing to understand the phenomenon.
Predictive HR Analytics

• Predictive HR analytics through statistical relationships can align HR actions with the achievement or
otherwise of the organizational strategic goals.

• Predictive HR analytics help HR managers evaluate the future implications of their decisions, and
accordingly develop long-term strategy.

• Key areas where HR can take decisions based on predictive modelling are:

1. Segmenting and profiling of employees

2. Analysis of employee loyalty and attrition
3. Analysis of recruitment needs
4. Employee sentiment analysis
5. Employee risk analysis
Present Scenario of HR Analytics
Importance and Significance of HR Analytics
• HR analytics provides scientific facts to HR.

• HR analytics start at the outset with the identification key business concerns, focus on strategy
and long-term sustainability issues. After this it assesses available knowledge and competencies-
set in the organizations, understand the gap, if any, and then focus on appropriate interventions.

• HR managers decide for appropriate interventions factoring adverse effect of decisional

outcomes. This is a working model of HR analytics.

• With HR analytics, HR managers can predict the trend, which help them to take wise decisions,
evaluating different decisional alternatives, based on the big-data-set.

• With the increased understanding of the business goals of the organizations, HR managers can
align their HR decisions which facilitate achieving the business results.

• By mapping the organizational strategies, HR managers can adapt their HR plans and
programmes, which can help in achieving the strategic intents and so on.
 Steps for Alignment of HR Analytics with
Business Goals and Strategies
• Alignment of HR analytics with business goals and strategies of organizations requires
adherence to certain steps.
•
1. Framing of queries or questions.
2. Understanding appropriate data and metrics.
3. Building appropriate platform for HR analytics.
4. Gradual enhancement of HR analytics capabilities.
5. Disseminating the importance and value of HR analytics.

• All the above steps are commonly followed in organizations for better alignment of HR
analytics with the strategies and business goals. However, depending on the nature of
the organization, it may vary.
Python for ‘HR Analytics’
Why Python for HR Analytics?
 Python has large and integrated collection of tools for data manipulation,
analysis and processing.
 It is developer-friendly and high level programming language.
 Python code is interpreted by interpreter.
 It is platform independent programming language
 Python is dynamically-typed.
 Python has an integrated suite of software facilities for graphical
interpretation.
 The rich library of Python is immensely helpful in advance algorithms
related to different specializations.
 The machine learning algorithms of Python has brought a revolution in the
field of Data Analytics.
1.4.2 Variables in Python

A Python identifier is a name which is used to identify a variable, function, or any other user-
defined item.

Python is a case-sensitive programming language.

It should be noted that Python is a dynamically typed language unlike any other programming
language; hence, there is no need to declare variables in Python before their usage in the
program. The name of an identifier must follow the naming rules.
Core Libraries in Python
Library Purpose
NumPy The most fundamental library around which the scientific computation stack is
built is “NumPy” (Numerical Python).
pandas A Python library designed to work with labeled and relational data is “pandas”.
Matplotlib A Python library that is tailored for creating simple and powerful visualizations
with ease is “Matplotlib”.
seaborn The “seaborn” library is mostly focused on the visualization of statistical models;
such visualizations include heat maps and those that summarize data but still
depict overall distributions.
SciPy The “SciPy” library contains modules for linear algebra, optimization, integration,
and statistics.
statsmodels The “statsmodels” library for Python enables its users to conduct data
exploration via the use of various methods of estimation of statistical models
and perform statistical assertions and analysis.
 Employee Data Exploration
Using

Python Core Modules and Libraries

Employee Data Exploration

• Organizations have a vast amount of valuable data related to workforce, which the human
resources department takes great care of.

• It is important for the HR professional to understand the data before taking any decision or
performing advanced analysis.

• Thus, employee data exploration starts from organizing data into a highly usable and valuable
resource and extracting information to improve the functioning of the human resources
department.
 Overview
 In the last decades, having the best machines was enough to be competitive or to
dominate an industrial sector.

 Nowadays, the company that has more engaged and productive employees will have a
better chance of winning market competition.

 For this reason, companies can not lose important employees and when that begins to
happen you need to understand why, to prevent this from happening.

 The Human Resources Analytics dataset, is used to explain the first steps in the data
analysis path. In this first part is presented how to get familiarize itself with the data set by
performing the descriptive analysis.

 Techniques such as Exploratory Data Analysis (EDA) allow us to present the data in a more
meaningful way, applying general statistical methods and exploratory graphics, that allow a
simpler interpretation before engage a machine learning algorithm.
Exploratory Data Analysis(EDA)
Exploratory data analysis employs a variety of techniques (mostly statistical graphics) before
making inferences from data. It is essential to examine all variables in the dataset to:

 Catch mistakes
 Generate hypotheses
 See patterns in the data
 Extract important variables
 Detect outliers and anomalies
 Gain deep familiarity with the dataset
 Refine selection of features that will be used to build the machine learning models.
Boxplot Visualization
Boxplot
• A boxplot is a way of summarizing a set of data measured on an interval scale. It is often
used in exploratory data analysis. It is a type of graph which is used to show the shape of
the distribution, its central value, and variability.

• A boxplot displays the values of the most extreme values in the data set (maximum and
minimum values), the lower and upper quartiles, and the median.

• Boxplot is a good statistical graphic to analyze the dataset and identify outliers values.

• An outlier is as observation that lies an abnormal distance from other values, in this case
the analyst have to decide what is considered abnormal.
Correlation

• The correlation is a very useful statistical analysis that describes the degree of relationship
between two variables.

• Correlation measures the strength of association between the two variables.

Performance Analysis

• Usually there are 2 distinct groups of employees. A group with poor performance and other
with high performance employees.

• It's natural that employees that don't work well leave the company, but the main problem
is that the high performance employees is leaving too and it's necessary to understand
why.
Working hours

• Again, there are 2 groups of employees. A group that works fewer hours and another that
works more hours compared to the average hours worked.
 Summary of the Exploratory Data Analysis
1. It is a relatively young company, on average, employees have 3 or 4 years in the company and the
oldest employees are working 10 years.

2. The biggest difference in the salary from who stayed and those who left, was found in the
management department, in the others departments although the salaries of who stayed be
higher in average, it is not a big difference.

3. The number of employees that had a work accident is about 14%, of which only 169 employees left
the company, so don't seem to have a correlation with the employees leaving.

4. In five years only 2% of the employees were promoted. Is possible that many employees get
unmotivated and start planning to leave.

5. Employees with 7 or longer in the company didn't left. Employees with 5 years have more chances
to leaving.
6. There are 2 distinct groups of employees performance that left. A group with poor performance
with 2 projects and others with high performance with 5 or more projects. It is not necessary
retain all the employees, the focus is on keeping employees with high performance.

7. The employees with 4 years in the company have the lowest average satisfaction level of all the
company with (0.47).

8. The satisfaction drops when the employees are working in 5 or more projects. A number of 3 or
4 projects seems to be ideal independent of the time spend in the company.

9. The employees with 5 or more projects that left also worked at least 20% more hours than the
average of the company.

10. The satisfaction level of the employees that left is grouped in totally disappointed, below the
average satisfaction and satisfied.
Happy Analyzing