Course - Time Series Analysis
https://365datascience.teachable.com/p/time-series-analysis-in-python
https://365datascience.teachable.com/courses/enrolled/666338
https://www.toptal.com/artificial-intelligence/retail-sales-forecasting-ai
Introduction to Time Series in Python
- Start Introduction to Time Series Data
- Start Notation for Time Series Data
- Start Peculiarities
- Start Loading the Data
- Start Examining the Data
- Start Plotting the Data
- Start The QQ Plot
Creating a Time Series Object in Python
- Start Transforming String inputs into DateTime Values
- Start Using Dates as Indices
- Start Setting the Frequency
- Start Filling Missing Values
- Start Adding and Removing Columns in a Data Frame
- Start Splitting up the Data
- Start Appendix: Updating the Dataset
Working with Time Series in Python
- Start White Noise
- Start Random Walk
- Start Stationarity
- Start Determining Weak Form Stationarity
- Start Seasonality
- Start Correlation Between Past and Present Values
- Start The ACF
- Start The PACF
Picking the Correct Model
Start A Quick Guide to Picking the Correct Model
The Autoregressive Model
- Start The AR Model
- Start Examining the ACF and PACF of Prices
- Start Fitting an AR(1) Model for Index Prices
- Start Fitting Higher Lag AR Models for Prices
- Start Using Returns
- Start Examining the ACF and PACF of Returns
- Start Fitting an AR(1) Model for Index Returns
- Start Fitting Higher Lag AR Models for Returns
- Start Normalizing Values
- Start Model Selection for Normalized Returns
- Start Examining the AR Model Residuals
- Start Unexpected Shocks from Past Periods
The Moving Average (MA) Model
- Start The MA Model
- Start Fitting an MA(1) Model for Returns
- Start Fitting Higher-Lag MA Models for Returns
- Start Examining the MA Model Residuals for Returns
- Start Model Selection for Normalized Returns
- Start Fitting an MA(1) Model for Prices
- Start Past Values and Past Errors
The Autoregressive Moving Average (ARMA) Model
- Start The ARMA Model
- Start Fitting a Simple ARMA Model for Returns
- Start Fitting a Higher-Lag ARMA Model for Returns - part 1
- Start Fitting a Higher-Lag ARMA Model for Returns - part 2
- Start Fitting a Higher-Lag ARMA Model for Returns - part 3
- Start Examining the ARMA Model Residuals of Returns
- Start ARMA for Prices
- Start ARMA Models and Non-stationary Data
The Autoregressive Integrated Moving Average (ARIMA) Model
- Start The ARIMA Model
- Start Fitting a Simple ARIMA Model for Prices
- Start Fitting a Higher Lag ARIMA Model for Prices - part 1
- Start Fitting a Higher Lag ARIMA Model for Prices - part 2
- Start Higher Levels of Integration
- Start Using ARIMA Models for Returns
- Start Outside Factors and the ARIMAX Model
- Start Seasonal Models - the SARIMAX Model
- Start Predicting Stability
The ARCH Model
- Start The ARCH Model
- Start Volatility
- Start A More Detailed Look of the ARCH Model
- Start The arch_model Method
- Start The Simple ARCH Model
- Start Higher Lag ARCH Models
- Start An ARMA Equivalent of the ARCH Model
The GARCH Model
- Start The GARCH Model
- Start The ARMA and the GARCH
- Start The Simple GARCH Model
- Start Higher-Lag GARCH Models
- Start An Alternative to the Model Selection Process
Auto ARIMA
- Start Auto ARIMA
- Start Preparing Python for Model Selection
- Start The Default Best Fit
- Start Basic Auto ARIMA Arguments
- Start Advanced Auto ARIMA Arguments
- Start The Goal Behind Modeling
Forecasting
- Start Introduction to Forecasting
- Start Simple Forecasting (Returns with AR and MA)
- Start Intermediate Forecasting (MAX Models)
- Start Advanced Forecasting (Seasonal Models)
- Start Auto ARIMA Forecasting
- Start Pitfalls of Forecasting
- Start Forecasting Volatility
- Start Appendix: Multiple Regression Forecasting
Business Case
Start Business Case - A Look Into the Automobile Industry
Time Series Forecasting on Stock Prices
Time Series Made Easy in Python - darts documentation
darts is a Python library for easy manipulation and forecasting of time series. It contains a variety of models, from classics such as ARIMA to deep neural networks. The models can all be used in the same way, usingfit()andpredict()functions, similar to scikit-learn. The library also makes it easy to backtest models, and combine the predictions of several models and external regressors. Darts supports both univariate and multivariate time series and models. The ML-based models can be trained on multiple time series, and some of the models offer probabilistic forecasts