Time Series Workshop
Quant
Introduction to Time Series
Overview
I’m hosting a Time Series Workshop for Fuqua’s MQM and MSQM programs.
- 25Mar2026 from 4pm-6pm in RJR
- Live 2hr session
- Focus on Macro & Finance applications
- Data Prep, 2) Explore, 3) Explain, 4) Forecast
New additions to the workshop
I’m lucky enough to host this talk at least once a year. Some new additions:
- Global Models (e.g. deepAR)
- Foundation Models (e.g. Chronos)
- Uncertainty quantification via Conformal methods
Slides
- I’ve prepared a relative expansive slide deck, which you can use as a reference guide.
- We will not cover every slide/topic during the live session.
- Instead, we will eschew the details in favor of a quick overview. You can dive deeper later.
Case Study
A Few Useful Time Series packages
R Packages
1. Data Preparation
Core Data Handling
- data.table
- dplyr
- tidyr
Time Index & Date Handling
- lubridate
- zoo
- xts
- tsibble
Missing Data & Interpolation
- imputeTS
- zoo (na.approx)
Mixed Frequency / Alignment
- tempdisagg
- midasr
Data Sources
- quantmod
- tidyquant
- fredr
2. Explore
Autocorrelation / Cross-correlation
- stats (acf, pacf, ccf)
- forecast
- TSA
Stationarity & Unit Root Testing
- tseries
- urca
- fUnitRoots
Cointegration
- urca
- dynlm
- tsDyn
Clustering Time Series
- TSclust
- dtw
- cluster
Decomposition / Filtering / Seasonality
- forecast
- stats (stl)
- seasonal (X-13 ARIMA-SEATS)
- mFilter
- feasts
Dimension Reduction
- stats (prcomp)
- FactoMineR
- dynfactoR
- dfms
- keras
- torch
3. Explain (Modeling)
Linear / Econometric Models
- dynlm
- lm
- plm
ARIMA / SARIMA
- forecast
- fable
VAR / Multivariate Models
- vars
- tsDyn
Error Correction Models (ECM)
- dynlm
- urca
- tsDyn
State Space Models
- KFAS
- dlm
- bsts
Volatility Models
- rugarch
- fGarch
4. Forecast
Classical Forecasting
- forecast
- fable
- prophet
Forecast Evaluation
- yardstick
- MLmetrics
Machine Learning
- caret
- tidymodels
- randomForest
- xgboost
Deep Learning / Global Models
- torch
- keras
- modeltime
Foundation Models (via Python)
- reticulate (interface to Chronos, TimeGPT)
Python Packages
1. Data Preparation
- pandas
- numpy
- polars
- dateutil
- fancyimpute
- midaspy
- yfinance
- pandas_datareader
- fredapi
2. Explore
Autocorrelation / Cross-correlation
- statsmodels
Stationarity
- statsmodels
- arch
Cointegration
- statsmodels
- arch
Clustering
- tslearn
- dtaidistance
- scikit-learn
Decomposition / Filtering
- statsmodels
- pmdarima
- x13_arima_analysis
Dimension Reduction
- scikit-learn
- statsmodels
- tensorflow
- pytorch
3. Explain (Modeling)
Linear / Econometric
- statsmodels
- linearmodels
ARIMA / SARIMA
- statsmodels
- pmdarima
VAR / VECM
- statsmodels
ECM / Cointegration
- statsmodels
State Space Models
- statsmodels
- pykalman
Volatility Models
- arch
4. Forecast
Classical Forecasting
- statsmodels
- pmdarima
- prophet
Machine Learning
- scikit-learn
- xgboost
- lightgbm
Deep Learning / Global Models
- gluonts
- pytorch-forecasting
- darts
Foundation Models
- nixtla (TimeGPT)
- amazon-chronos
- transformers