Asit Dave

Stock price forecasting web-app and analysis using EDA, technical indicators, portfolio risk assessment, and time-series modeling (ARIMA & LSTM).

Causal analysis of advertisement effectiveness using A/B testing, Bayesian inference, and a cloud-deployed analytics dashboard.

This repository aims at understanding the cosmic web through T-Web classification scheme. It calculates the Tidal fields in a cosmological simulation box and uses it to classify the large-scale structures of the Universe.

Segments app users by behavior and predicts uninstall likelihood using ML models to support data-driven retention strategy.

Analyzes hourly web traffic and forecasted engagement to guide content strategy.

A collection of data science projects that explore various datasets, applying analysis, machine learning, and data visualization techniques to extract valuable insights and solve real-world problems.

Creates a deep learning workflow to classify fashion clothing images using PyTorch

Classifies aircraft surface damage (dent vs crack) using VGG16 and generates descriptive damage captions with BLIP for intelligent maintenance analysis.

Detects suspicious financial transactions using data analysis and an Isolation Forest model, with an interactive tool to flag potential anomalies.

Analyzes food delivery order data to uncover cost drivers, evaluate discount and commission strategies, and model ways to improve unit profitability.

Analyzes retail sales and profit data to uncover trends, top products, regional performance, and the impact of discounts using Python and interactive visualizations.

Predicts Falcon 9 landing success and estimated launch cost using real mission data and machine learning

Quantitative analysis of option pricing, volatility modeling, and portfolio risk using the Black-Scholes model and Monte Carlo simulation

No description provided.

Predicting wine quality from physicochemical properties using Random Forest model with an interactive Streamlit app.

Data-driven analysis of marketing channel performance other than evaluate CAC, ROI, and campaign efficiency.

Used SQL to answer business questions from sales data

GitHub portfolio page

No description provided.

No description provided.

A portfolio of machine learning projects that use classification, clustering, and regression to analyze data, build predictive models, and extract valuable insights for real-world applications.

BIPOLARS is a Python-based pipeline for computing higher-order statistical measures, such as bispectrum and power spectrum multipoles, from galaxy survey data, optimized for parallel processing and cosmological analysis.

This python script simulates the hierarchical merging of black holes for different star cluster environments: Open, Globular, and Nuclear.