# Schedule

*Italics* indicate that laptops are required for lab activities.

- August 31
- Lecture: The forecasting workflow
*Lab:*Set up`R`

, Intro to R assignment

- Sept. 2
- Discussion: Why dynamics and forecasting?
- Reading: Clark et al. 2001
- Reading: Houlahan et al. 2017

*Lab:*Intro to R assignment

- Discussion: Why dynamics and forecasting?

- September 7
- DUE: Before class install the latest version of R, the latest version of RStudio, and these R Packages: forecast, ggplot2, lubridate, dplyr, scales, gridExtra, ggthemes, reshape2, zoo
- Reading: Look through the following NEON Working with Time Series Data Tutorials
- Time Series 02: Date-time conversions
- Time Series 04: Manipulate data with dplyr
- Time Series 05: Plot time series with ggplot2
- Also, Julian day conversion

- Reading: Look through the following NEON Working with Time Series Data Tutorials
*Lab:*Working with time series data in R- Download data for the lab
- Optional Weecology video lecture: Dates/times

- DUE: Before class install the latest version of R, the latest version of RStudio, and these R Packages: forecast, ggplot2, lubridate, dplyr, scales, gridExtra, ggthemes, reshape2, zoo

- September 9
*Lab:*Time series decomposition and autocorrelation assignment- Optional Weecology video lecture: Time series decomposition
- Optional Weecology video lecture: Time series autocorrelation

- September 14
- DUE: Time series decomposition and autocorrelation assignment
- Lecture: Introduction to time series modeling demo code
- Optional Weecology video lecture: ARIMA

- September 16
- Discussion: Introduction to forecasting
*Lab:*- Introduction to forecasting demo code
- Introduction to forecasting assignment
- Optional Weecology video lecture: Intro to forecasting

- September 21
- Lecture: Time series and population models
*Lab:*

- September 23
- DUE: Introduction to forecasting assignment
- Discussion: Understanding vs. prediction
- Reading: Breiman 2001(skim the examples)

- Discussion: Understanding vs. prediction and model selection
- Lecture: Bias-variance trade-off

- September 28
- Discussion: The importance of uncertainty
- Reading: Dietze Chapter 2

*Lab:*- Simulating prediction intervals
- If time permits:
- Forecast evaluation demo code
- Begin Evaluating forecasts assignment
- Optional Weecoogy video lecture: Forecast evaluation

- Discussion: The importance of uncertainty

- September 30
*Lab:*

- October 5
- Presentation: Weather forecasting (Peter)
*Lab:*- Work on Evaluating forecasts assignment

- October 7
- DUE at the end of this class: Evaluating forecasts assignment
- Presentation: Economic forecasting (Silver chapter on Canvas) (Jack)
*Lab:*Begin Forecasting Challenge #1

- October 12
- Lecture: Bias-variance trade-off
*Lab:*Work on Forecasting Challenge #1

- October 14
- Presentation: Bias-variance trade-off
*Lab:*Work on Forecasting Challenge #1

- October 19
- Presentation: Forecasting phenology (Michael)
*Lab:*Work on Forecasting Challenge #1

- October 21
- Presentation: Population dynamics (Dani), perhaps animals or plants
*Lab:*Work on Forecasting Challenge #1- Example code for regularization

- October 26
- Presentation: Veronica: Invasions
*Lab:*Work on Forecasting Challenge #1

- October 28
- DUE at 9:00 am: Forecasting Challenge #1
- Scores calculated, winners announced!
- Debrief the challenge

- Nov. 2
- DUE: Read up on NEON Forecast Challenge
- Decide as a group on next forecasting challenge. NEON?
- Soren on Political forecasting

- Nov. 4
- Courtney on epidemics
- DUE: Read Intro to Bayes
- Lecture: Bayesian modeling: background

- Nov. 9
- Erika on carbon cycle
- Lecture: Bayesian modeling: in practice

- Nov. 11
- John on SDM validation
- DUE: 1) Install the JAGS library; 2) Install the jagsUI R package; 3) run the JAGS example in Bayesian modeling: in practice; 4) add an additional climate covariate and refit the model.
- Lecture: Bayesian modeling: in practice continued

- Nov. 16
- DUE Ethics of forecasting: Start with Record et al., a blog post, and then read Hobday et al.
- DUE try to install Stan (
`rstan`

package) and run the example code in the lecture below - Lecture: Bayesian modeling: in practice

- Nov. 18
- DUE:
- Explore flu data and R package
- Explore Bayesian time series analysis in R with bayesforecast, example code
- Review flu forecasting literature

*Lab*: Start working on flu forecast!

- DUE:

- Nov. 23
- DUE: Read Things to know… and this review.
- Dani is working on Bayes time series code, Erika will make time series figures for the flu data, and I will make some decisions about forecast targets, and some others(?) are looking for COVID data proxies…
- Here is a Google Drive folder containing data and code. Feel free to download and play with scripts, or add your own to the drive, just be careful about deleting things…

- Nov. 25 NO CLASS (Thanksgiving)

- Nov. 30
- TO DO:
- Michael: Hierarchical time series models
- Soren: COVID covariate
- Erika: compare seasonal components by region
- Courtney: NAs?
- Jack: Flu vaccination data?
- Peter and Dani: Why are the Bayes time series models making bad forecasts?

- TO DO:

- Dec. 2

- Dec. 7

- December 9
- Wrap-up discussion: Can we forecast in ecology (and what can we forecast)?
- Discussion questions

- OLD LINKS
- Population forecasts – animals
- Population forecasts – plants
- Biodiversity forecasting
- Epidemiological forecasts, like State-space modeling to support management of brucellosis in the Yellowstone bison population
- Carbon cycle
- Scenario Planning: a Tool for Conservation in an Uncertain World
- Use of remote-sensing in forecasting
- Validation of species-distribution models? Maguire et al. ; Blois et al.