Palash Halder

Data Analyst

80 · Excellent

About

Final-year B.Tech CSE student and aspiring Data Analyst with hands-on experience turning raw data into decision-ready insights - SQL querying, Exploratory Data Analysis (EDA), KPI reporting, and interactive dashboards (Power BI, Streamlit) across 4+ deployed projects. Published peer-reviewed researcher (IJCRT) applying ML architectures (XGBoost, GNNs, VAEs) to computational chemistry. Analyzed 6 years of national-scale transaction data, quantified multi-year growth trends, and forecasted future volumes using time-series modeling and statistical analysis. Strong foundation in Python, SQL, Excel, Pandas, and data visualization/BI tools, with production deployments and dashboards teams can inspect live rather than just read about. Seeking a Data Analyst role to translate data into actionable business insights from day one.

Skills & Expertise (47)

Machine Learning Advanced

8.4/10

Years Exp

Exploratory data analysis Advanced

8.0/10

Years Exp

SQL Advanced

8.0/10

Years Exp

Data Cleaning Advanced

8.0/10

Years Exp

Python Advanced

8.0/10

Years Exp

Power BI Intermediate

7.6/10

Years Exp

Business Intelligence Intermediate

7.6/10

Years Exp

Feature Engineering Intermediate

7.4/10

Years Exp

Deep Learning Intermediate

7.4/10

Years Exp

KPI Reporting Intermediate

7.0/10

Years Exp

Voice Recognition Supervised Learning Unsupervised Learning Classification Regression Postman XGBoost Time-Series Forecasting NLP Fundamentals Face Recognition SQLite Supabase Jupyter Notebook Vercel Render Streamlit Cloud Git GitHub JavaScript Hypothesis Testing Statistical Analysis Data Wrangling Data Storytelling Dashboarding MS Excel Java HTML CSS Streamlit Pandas NumPy scikit-learn Matplotlib Seaborn Plotly Facebook Prophet Flask

Work Experience

Data Analyst

PhonePe Pulse

Present - Present

Analyzed 6 years of national UPI transaction data end-to-end, quantifying 40x growth in transaction volume (2018–2024) to surface strategic market trends. Authored 7 advanced SQL queries (CTEs, LAG() window functions, RANK() OVER PARTITION BY, multi-table joins) to answer business questions on growth, churn risk, and market concentration. Built a Facebook Prophet forecasting model with multiplicative seasonality, projecting national transaction volume 4 quarters ahead with confidence intervals. Translated findings into merchant-acquisition recommendations segmented by state-level growth and engagement across Maharashtra, Karnataka, and Telangana - deployed live on Render.

Machine Learning Engineer

Employee Attrition Prediction System

Present - Present

Built an end-to-end ML application predicting attrition risk from 30+ workforce features, giving HR teams early visibility into flight-risk employees. Ran full EDA and feature engineering - handled class imbalance, encoded categoricals, and ranked predictors by correlation and feature-importance score. Optimized a classification model for precision-recall trade-off and serialized it with Joblib for production use. Shipped a dual-interface Streamlit app (individual prediction + workforce analytics dashboard), live on Streamlit Cloud.

AI Engineer

SnapClass

Present - Present

Engineered a multimodal system combining face recognition (Dlib) and voice recognition (Resemblyzer) to automate classroom attendance with zero manual input. Built pipelines for face-embedding generation, multi-face detection from group photos, and speaker identification at scale. Designed a Supabase (PostgreSQL) backend for student records and attendance, with a dual-portal Streamlit app for teachers and students - deployed with a dedicated landing page on Vercel.

Full-Stack Developer

DownZORA

Present - Present

Built a full-stack Flask app to fetch, validate, and process publicly available media links, with REST APIs for client-server communication and a fully responsive frontend.