HR Analytics - Mercedes Benz

Successfully analyzed data collected by the HR department and constructed a robust predictive ML model that accurately forecast an employee's likelihood of leaving the company.

About Project

Objective was to identify core factors that play a role in employees choosing to leave the company by analyzing employee retintion and constructing a Machine Learning model for accurately predicting employee attrition

Project was executed in following stages

Stage 1: Import Data & Packages
Stage 2: Data Exploration - EDA & Data Cleaning
Stage 3: Building Models
- Approach A - Building Regression Model
- Approach B - Building Tree Based Models( Decision Tree & Random Forest )
Stage 4: Model Evaluation & Sharing Insights

Project was executed in following stages

Stage 1: Import Data & Packages
Stage 2: Data Exploration - EDA & Data Cleaning
Stage 3: Building Models
- Approach A - Building Regression Model
- Approach B - Building Tree Based Models( Decision Tree & Random Forest )
Stage 4: Model Evaluation & Sharing Insights

Project was executed in following stages

Stage 1: Import Data & Packages
Stage 2: Data Exploration - EDA & Data Cleaning
Stage 3: Building Models
- Approach A - Building Regression Model
- Approach B - Building Tree Based Models( Decision Tree & Random Forest )
Stage 4: Model Evaluation & Sharing Insights

Project was executed in following stages

Stage 1: Import Data & Packages
Stage 2: Data Exploration - EDA & Data Cleaning
Stage 3: Building Models
- Approach A - Building Regression Model
- Approach B - Building Tree Based Models( Decision Tree & Random Forest )
Stage 4: Model Evaluation & Sharing Insights

Project Stack

Python, EDA, Machine Learning

Project Stack

Python, EDA, Machine Learning

Role

Data Analyst, Data Scientist

Role

Data Analyst, Data Scientist

Services

EDA, Data Viz, Processing, ML Models

Services

EDA, Data Viz, Processing, ML Models

CodeBase

GitHub Repo

CodeBase

GitHub Repo

Jupyter Notebook

Nbviewer

Jupyter Notebook

Nbviewer

"I began with data import and conducted exploratory data analysis (EDA), followed by a two-sample hypothesis test. I then built Random Forest and XGBoost models, improved their scores with model tuning, and selected XGBoost as the champion model for predicting user churn!"

Waqas Ahmad – Data Analyst

Take me back home