Detect U.S. housing market bubbles using macroeconomic signals. Forecast HPI, score speculative risk, and visualize insights using a fully modular, cloud-native GCP pipeline.

Time Series Forecast of traffic on JetRail

Forecast for Bitcoin till April 2026 based on linear regression and std devs

Market Forecasting Model powered by Fastapi. The model uses garch model to generate forecast for market volatility in India.

Dual-domain AI framework for airline RM and digital analytics validation

Recurrent neural networks to predict solar radiation measurements.

This project analyzes transaction data and generates forecasts using machine learning models like ARIMA, Prophet, and others. It outputs forecasted data, visualizations, and saved models for further analysis.

This multihead attention transformer model is trained for Nelson Wind Speed 24 hours ahead 10-minute interval forecast.

Forecasting USD/TRY exchange rates using LSTM (Deep Learning) and XGBoost (Machine Learning). Trained on historical Yahoo Finance data (2018–2024) to predict the next day’s closing rate.

Forecasting bed linen sales and seasonality trends for Mungo using fictional data (ARIMA forecasting)

Time series forecasting of passport demand for a government entity (non-disclosure agreement)

Light pollution analyzer - forecasting, clustering, find key factor influencing light pollution, outliers state in datasheet all with python and Google earth engine.

This repository is a quick project to transfer my TensorFlow framework knowledge to Pytorch

Build predictive analytics solution for inventory management.

The purpose of our project is to propose an uncertainty model for stock price forecasting under non-normality. The data used for this analysis is from the Shanghai and Shenzhen stock exchanges.

Advanced time series forecasting with transformer architectures that can predict complex patterns across multiple horizons with uncertainty quantification.

This repository uses machine learning models like Random Forest, XGBoost, LightGBM, and time-series forecasting with Prophet to predict game search volumes. Additionally, Grid Search is applied for hyperparameter tuning of the LightGBM model.

LLM-Forecast: A Novel Hybrid Forecasting Methodology Integrating ARIMA and Large Language Models

Implementation of naive time-series forecasting models

In this repository, I have basically analyzed the S&P 500 stocks data and found various meaningful insights in the data.