Portfolio - Hritik kumar | Codebasics

Powered by
Codebasics AI Engineering & DS Bootcamp

Hritik kumar

Data Aspirant

EMAIL

[email protected]

PHONE

+91 9546875952

Download CV

About Me

I'm Hritik Kumar, a Data Science graduate who turns messy data into decisions. I work across the full analytics lifecycle — cleaning data, building predictive models with Python and Scikit-learn, writing SQL to uncover trends, and designing Power BI/Tableau dashboards that stakeholders actually use.
I've predicted health insurance premiums using regression models, profiled banking customer behavior to drive credit-risk decisions, and built a full-stack expense tracker with FastAPI and Streamlit. What I bring isn't just technical skill — it's the ability to take ambiguous problems and turn them into clear, business-ready answers.
Currently looking to bring that mix of analytics, automation, and clear communication to a team solving real operational problems — especially in transport, infrastructure, or financial services.

1

Python Project

2

Machine Learning Projects

Key Skills

My Projects

Awards & Certificate

Python: Beginner to Advanced For Data Professionals

SQL for Data Science

Math and Statistics For AI, Data Science

Master Machine Learning for Data Science & AI: Beginner to Advanced

Natural Language Processing

Gen AI to Agentic AI with Business Projects

Let's Work Together!

Feel free to get in touch with me. I am always open to discussing new projects, creative ideas or opportunities to be part of your visions.

EMAIL

[email protected]

Call Me

+91 9546875952

LinkedIn

https://www.linkedin.com/in/hritik-kumar-865236349?utm_source=share&utm_campaign=share_via&utm_content=profile&utm_medium=android_app

GitHub

https://github.com/Hritikv16

Python Project

Bollywood vs Hollywood — Box Office & Ratings Analysis

Domain/Function: Media & Entertainment (specifically: Film Industry Analytics)

Analyzed Bollywood vs Hollywood movies — comparing budgets, revenue, ROI, and IMDb ratings — using Python, Pandas, NumPy, and Matplotlib. Cleaned multi-currency, multi-unit data to uncover profitability trends and explore how spend relates to audience reception across both film industries.

View Project
on GitHUB

Bollywood vs Hollywood — Box Office & Ratings Analysis

Bar chart comparing avg budget vs revenue (USD millions) across Bollywood & Hollywood. Currencies and units were standardized first — converting INR/USD and thousands/millions/billions into one consistent scale for accurate comparison.

Bollywood vs Hollywood — Box Office & Ratings Analysis

Scatter plot: budget vs IMDb rating by industry. Shows no strong correlation — lower-budget films often outperform big-budget ones, especially in Bollywood. Storytelling matters more than spend.

Project Details

Project Title Options — Bollywood vs Hollywood Movie Analysis
Resume-style: "Bollywood vs Hollywood — Box Office & Ratings Analysis," "Movie Industry Analytics — Budget, Revenue & Rating Trends," or "Global Film Industry Performance Analysis." These are clear, professional, and instantly tell a recruiter what the project does.
Portfolio-style: "CineMetrics: Comparing Bollywood & Hollywood Box Office Performance," "Lights, Camera, Data: A Bollywood vs Hollywood Analytics Study," or "Box Office Breakdown: Budget-to-Revenue Analysis Across Film Industries." These add a bit more personality for a GitHub or portfolio audience.
Recommended title: "Bollywood vs Hollywood — Box Office & Ratings Analysis." It stays consistent with the tone of other resume projects like "ML Health Premium Prediction" and "Atliqo Bank Analysis," making it easy for recruiters to scan and understand at a glance while still signaling clear business relevance in media analytics.

Machine Learning Project

Insurance premium prediction project

Domain/Function: insurance premium prediction project

Built an insurance premium prediction pipeline using Python, Pandas & NumPy for data cleaning, Seaborn/Matplotlib for EDA, Statsmodels for multicollinearity checks, and Scikit-learn (Linear/Ridge/Lasso) plus XGBoost for regression modeling with hyperparameter tuning.

View Project
on GitHUB

View Project
on Github

Other Tools

Python

Pandas

NumPy

Matplotlib

Seaborn

Scikit-learn

Statsmodels

XGBoost

Insurance premium prediction project

Correlation heatmap revealing feature relationships and multicollinearity, guiding feature selection before regression modeling

Insurance premium prediction project

XGBoost feature importance ranking age, plan type, and BMI as top drivers of predicted insurance premiums

Project Details

Built an end-to-end machine learning pipeline to predict health insurance premiums based on customer demographic and lifestyle attributes such as age, BMI, smoking status, number of dependants, and medical history. The workflow began with data cleaning and preprocessing in Pandas and NumPy — handling missing values, duplicates, and invalid entries (e.g., negative dependants, unrealistic ages) — followed by exploratory data analysis using Seaborn and Matplotlib to uncover outliers and feature relationships. Multicollinearity among numerical features was assessed using Variance Inflation Factor (VIF) from Statsmodels. Multiple regression models (Linear Regression, Ridge, Lasso, and XGBoost) were trained and evaluated using R², MSE, and MAE, with hyperparameters optimized via GridSearchCV and RandomizedSearchCV. The final model provides accurate premium estimates, demonstrating practical application of data cleaning, feature engineering, and predictive modeling in the insurance domain.

Machine Learning Project

Energy Drink Price Range Prediction using Machine Learning

Domain/Function: Pricing Analytics / FMCG (Retail & Consumer Goods)

Built an ML model to predict the ideal price range for an energy drink based on customer demographics and buying behavior. Compared 6 models, tracked experiments with MLflow, and deployed the best (XGBoost, 91.8% accuracy) as an interactive Streamlit app

View Project
on GitHUB

Other Tools

Python

Pandas

Scikit-learn

XGBoost

LightGBM

MLflow

Streamlit

Plotly

Git/GitHub

DagsHub

Energy Drink Price Range Prediction using Machine Learning

"Interactive input form where users select a customer profile — demographics, consumption habits, and brand behavior — to generate a price prediction. Sidebar shows the model (XGBoost, 91.8% accuracy)."

Energy Drink Price Range Prediction using Machine Learning

Prediction result view showing the predicted price range with confidence score, plus a probability breakdown across all four price tiers — powered by the XGBoost model with 91.8% test accuracy

Project Details

AtliQ-style business problem: how should a company price an energy drink for different customer segments instead of guessing? I built an end-to-end ML pipeline using consumer survey data (demographics, consumption habits, brand behavior) to predict which price tier (₹50-100 to ₹200-250) a customer profile falls into.

I engineered features (label + one-hot encoding), trained and compared 6 classification models — Naive Bayes, Logistic Regression, SVM, Random Forest, XGBoost, LightGBM — and tracked every run's parameters, metrics, and model artifacts using MLflow for full experiment reproducibility. XGBoost was the best performer at 91.8% test accuracy.

To make it usable for non-technical stakeholders, I deployed the model as an interactive Streamlit app with a live prediction form, confidence chart, and model performance dashboard — built for real client-facing demos, not just a notebook. (≈960 characters)

Already have an account? Sign in

or