/human-resource-data-analysis-streamlit

A HR dashboard built on Streamlit that analyzes raw data to gain knowledge about how well organizational policies affect staff promotions and layoffs as well as employee behaviors like attrition and job satisfaction.

Primary LanguagePython

title emoji colorFrom colorTo sdk sdk_version app_file pinned license
Human Resource Data Analysis Streamlit
🏢
red
pink
streamlit
1.27.1
src/app.py
false
mit





Human Resource Analysis

In this Data Analysis project, we examine raw HR data to learn more about how organizational policies influence employee behaviors such as attrition and job-satisfaction plus staff promotions and layoffs. The company's human resources policies shall then be amended based on the insights obtained.

Features

⚡Multipage Interactive Dashboard
⚡Deployed on cloud [Report accessible on the web]
⚡Streamlit
⚡Plotly
⚡Pandas

Table of Contents

Introduction

  • Datamatrix-ml Inc. is one of the top financial research firms, specializing in cutting-edge financial market research. It has over 1400 employees.
  • The R&D department of the company creates the exclusive financial research reports and advice papers
  • The company's sales division is in charge of finding the customers for its research goods.
  • The HR department is actively gathering data through internal questionnaires, interviews, and HR employment records to evaluate the efficacy of the workforce given the human-centric aspect of the corporate work culture.

Requirements

As part of their yearly evaluation of the employees and HR policies, the HR department has requested us to undertake a confidential data analysis project in order to gain insights into its workforce management. The questions that the HR department is seeking answers to are listed below.....

Requirement ID For Whom Requirement Description
HR-DA01-REQ-1 Executive Committee How gender-balanced is our workforce today?
Should we concentrate on employing more employees of a certain gender in the upcoming fiscal year?
HR-DA01-REQ-2 Executive Committee Do you think our organization is young or aging?
Should we concentrate on hiring people in a specific age and gender group?
HR-DA01-REQ-3 Executive Committee How evenly distributed in terms of marital status is the work force?
HR-DA01-REQ-4 Executive Committee Does the gender distribution of employees at the departmental level match that of the entire company?
HR-DA01-REQ-5 Executive Committee Does the gender distribution of employees at the departmental level match that of the entire company?
HR-DA01-REQ-6 Executive Committee How is each department performing on average in terms of "overtime," "manager's employee retention," and compensation?
HR-DA01-REQ-7 Executive Committee Based on their job experience, are our employee distributions balanced?
Do we require targeted hiring for a specific spectrum of job experience?
Do employees wish to spend the majority of their careers working with our company?
HR-DA01-REQ-8 Rightsizing Team How well are we doing on our 10% promotion goal?
Are we at or below the maximum 5% rate of performance-based layoffs?
Are the promotion and layoff rates at the departmental level consistent with those at the corporate level?
HR-DA01-REQ-9 Hiring & Retention Team Do we have a higher-than-10% attrition rate?
What are the elements that cause attrition directly?
HR-DA01-REQ-10 Executive Committee
Rightsizing Team
Hiring & Retention Team		 Need ability to filter and drill-down into data by Gender, Department, Education, Job Role, Age and Work Experience

Table-1 : Requirements

Dataset

  • The HR department collects and maintains the dataset. The relevant employee-related data attributes are provided by HR as a CSV dataset, and.
  • The dataset is anticipated to be evaluated and updated yearly in the same manner before being fed into the suggested data-analysis solution.
  • The input_data folder in the project repository contains the raw dataset.
  • Along with raw data the HR departments has supplied the business-rules to be used to compute employee's promotion-rate, retrenchment-rate and attrition-rate
Data Item Rule
Promotion % Rate [Count of to-be-promoted] =
     (if [yrs-since-last-promotion] >= 10 yrs and [performance-rating] is > 2)
[promotion-rate] = [Count of to-be-promoted] / [total-employee]) * 100
Retrenchment % Rate [Count of to-be-retrenched] =
    (if [Not-to-be-promoted] and ((if [years-in-current-role] is between 3 to 10 years and [performance-rating] is > 2)
OR (If [years-in-current-role] is >= 10 years and [performance-rating] < 3))
[retrenchment-rate] = [Count of to-be-retrenched] / [total-employee]) * 100
Attrition % Rate [attrition-rate] = [Count of employee left] / [total-employee] * 100

Table-2 : Rate Computation Rules

Solution Approach

The solution uses the 'Streamlit' and 'plotly' libraries to construct a fully interactive web-based cloud-hosted report dashboard. In order to gain insights and provide answers to the queries specified in Table-1: Requirements, the report includes the pertinent plots and tables. By zooming in on and filtering out a certain plot, users can interact with plotly plots. Additionally, the solution offers a group of global data filters that users can utilize to filter data throughout the entire application. These filters also dynamically change the visuals to reflect the data that remains after filtering. Users now have the option to explore specific areas of interest in the dataset by cutting, dicing, and drilling down.

Requirement ID Solution ID Proposed Solution
HR-DA01-REQ-1
HR-DA01-REQ-2
HR-DA01-REQ-3
HR-DA01-REQ-4
HR-DA01-REQ-5
HR-DA01-REQ-6
HR-DA01-REQ-7		 HR-DA01-SOL-1 To assist in providing answers to queries in the appropriate requirements,
a dedicated dashboard with interactive visualizations at the summary level
will be constructed
HR-DA01-REQ-8
HR-DA01-REQ-9		 HR-DA01-SOL-2 To assist in providing answers to queries regarding respective
requirements, a dedicated dashboard comprising interactive
visualizations relating to promotions and layoffs (capacity) will be created.
HR-DA01-REQ-10 HR-DA01-SOL-3 To assist in addressing queries on pertinent requirements, a dedicated
dashboard including dynamic visuals relevant to attrition will be constructed.

Exploratory Data Analysis (EDA) and Feature Engineering [pandas]:

The EDA and feature-engineering have been done using pandas to comprehend, become comfortable with, and verify the sanity of the given data. Using the pandas Python library, the data-exploration, data-cleaning, and feature-engineering have been completed. We are generally checking...

  • Presence of any missing values
  • Any unusual value (outliers)
  • Incorrect values (e.g., sales column, we see -ve numbers)
  • Determine categorical and numeric columns
  • Determine dimensions of categorical columns and range of numeric columns
  • Create new features required to show relevant data/numbers in the report

Report Creation [Streamlit, pandas, plotly]:

To put the suggested idea into practice, three interactive report dashboards (report tabs) will be developed. For specific requirements and the related proposed solution, see Table-3: Proposed Solution.
There are several graphics on each report dashboard, divided into expandable and collapsible sections. Each section provides the queries we're attempting to address along with any pertinent insights gained from the illustrations.

1. Executive Summary Dashboard [HR-DA01-SOL-1]:

'Gender', 'Age', 'Marital Status', 'Department', and 'Work Experience' are all examined in this high-level summary dashboard. This dashboard includes the sections below...

Gender & Age and Marital Status Statistics:



Department & Work Experience:



2. Capacity Dashboard [HR-DA01-SOL-2]:

This dashboard emphasizes the size and capabilities of the company's staff. This analysis will aid the business in its "rightsizing" decision in near future. This dashboard includes the sections below...

Summary level & Department level promotion and layoffs statistics:



3. Attrition Dashboard [HR-DA01-SOL-3]:

This dashboard looks at the company's capacity for employee retention and determines what variables are at play if attrition rates are higher than anticipated. The corporation may change its practices to better manage attrition in light of the findings.

How To Use

Read-only direct access via the web (Recommended):


and explore the fully functional interactive report.

Run the application locally:

To run (and modify) the application locally follow below steps...

  1. Clone this repo
  2. From within the root folder of the application install all dependencies by running 
    pip install requirements.txt
  3. From within the root folder of the application run the application... 
    streamlit run src/app.py
  4. Copy and paste the localhost:<port-no> web-address displayed by above command in your browser to open and application.

License

MIT License

Contact Me

email twitter linkedin website

Back To The Top