The practitioner's forecasting library

A repository for data visualization and data analysis (Occasionally ML) related projects.

R code for exchange rate prediction using Multilayer Perceptron (MLP) models with various architectures and evaluation metrics

This project aims to forecast the future web traffic for approximately 145,000 Wikipedia articles.

Project compares three regression models for predicting the amount of gold recovered from gold ore in order to optimize gold production and eliminate unprofitable parameters. Data provided by Zyfra.

Panel time-series forecasting notebooks (daily sales across stores × items). Clean validation (holdout + rolling-origin backtest), strong statistical baselines (SARIMAX/TBATS/ARIMA), and automated models (AutoTS), with optional Prophet/Darts/NeuralProphet. Primary metric: SMAPE.

Model to predict the amount of gold extracted from gold mineral.

데이콘 전력사용량 예측 AI 경진대회

Предсказание коэффициента восстановления золота из золотодобывающей руды

As a Data Scientist at Zyfra Company, I developed a predictive model to estimate gold extractable from ore. Using data on extraction and purification, I trained the model to enhance the efficiency of the production process.

2025 전력사용량 예측 경진대회: 시계열·기상·건물 메타데이터 결합 기반 SMAPE 최적화 파이프라인

[경진대회] 전력사용량 예측 AI 모델 개발

Energy consumption forecasting is crucial for efficient power management, grid stability, and energy resource planning. This project leverages Time Series Analysis techniques to predict energy consumption using GRU and LSTM models. By utilizing the PJM Interconnection LLC Energy Consumption Dataset, which records hourly power usage.

Integrated Project 2 from TripleTen

Zyfra is a pioneering developer of efficiency solutions for heavy industries & is aiming to take help of machine learning to optimize the efficiency in Gold Ore processing

📊 Explore panel time-series forecasting techniques for sales using popular Python libraries like ARIMA, Prophet, and AutoTS in Jupyter notebooks.