Introduction

This repository provides a high-level overview of modern techniques for time series analysis and Forecasting. Jupyter Notebooks will be used to explain concepts and show code and visualizations. I also developed an accompanying core library ts to provide useful functionality in a modular and systematic way to all notebooks.

Installation

You can install the core library by cloning this repository and then following these steps

git clone https://github.com/mleila/timeseries
cd timeseries
pip install -r requirements.txt

This will install the core library in your virtualenv and you'll be able to run the notebooks. Make sure you are using Python 3 (preferably 3.7).

Guide

The notebooks are divided into the following three categories

Time Series Concepts

These notebooks provide a high-level overview of some of the useful ideas for time series analysis and forecasting.

• Digital Signal Processing and Fourier Analysis
• Classic Time Series Models
• Machine Learning for Time Series Analysis and Forecasting
• Deep Learning for Time Series Analysis and Forecasting

Time Series Data Processing and Visualization

• Interesting Time Series Datasets
• Visualizing Time Series Data with Matplotlib and Plotly
• Spark and Pandas Time Series Pipelines
• Dealing with Missing Data

Popular Time Series Python Libraries

• Facebook's Prophet

• Statsmodels