- June 28, 2020
ENST90044 Analysing Ecosystems and Their Values May 17, 2020 Individual Modelling Projects THE PURPOSE OF THIS PROJECT is to provide an oppotunity to undertake some ecosystem-relevant modelling on a question and/or dataset that interest you. This will allow you to develop and apply some of the skills that you have acquired in this class over the past few months. In the previous handout I briefly described the four key components of the modelling project: 1. Develop a question related to some ecosystem value of interest to you. 2. Identify and obtain some relevant data 3. Develop a model to analyse the data 4. Assess your predictions in either time or space using withheld data In this handout, I’ll outline the main expectations for the project, including specific requirements for the final submission. THE PROJECT THAT YOU SUBMIT MUST contain all of the relevant and necessary information for me to assess the degree to which you have addressed each of the core components in the list above. This includes text, R code, model summary outputs, and relevant figures. The ability to convey science through compelling writing is an important skill. This means that you should make your writing as clear, vigorous, and engaging as possible. You have a story to tell, so tell it. For this assignemnt, I do not require or expect the standard IMRAD formatted report. The key written elements that you must include at a minimum are: • A description of your question and why you think it is interesting. • A description of your dataset(s) and where they came from. • A description of your model and how you choose it. • An assessment of your models and their predictive ability. • A short synthesis Research question: Your research question must be articulated in the first paragraph of the report. It should also be clearly referenced in your synthesis of the modelling project. ENST90044—Monitoring Ecosystems and Their Values 2 Datasets: You must describe in as much detail as possible what data you are using, where they are from (who collected them, where are they lodged, how did you access them), and why they are appropriate to answering your modelling question. Model: A description of your models. It is critical that you consider at least three different models in developing your project. These may be different statistical models (ie, inclusion of different predictor variables) or theoretical models (ie, different functional forms to express a relationship between or amongst predictors and response variables, or a combination of both. Predictive ability: Using the available data, develop your model using a subset of the data. Then assess the model by predicting the withheld data and comparing the predictions to withheld observations. Synthesis: Were you able to answer your question? Provide a short synthesis of what you found. Highlight where any bottlenecks were in developing your answer (ie, question, data, model, assessment). BASED ON PREVIOUS EXPERIENCES THERE are some important rules that you must follow and I will enforce. These should help you focus on the core story in your project. Word count/page limit: There is no formal word limit on this project. In the past, the final reports have been between 10-20 pages depending on the amount of code and figures included. Of that typically about 3-4 pages is written text. Figure limit: You may have up to (and including) 7 figures in your final report. You are welcome to use multi-panel figures, as appropriate. R code: You should show all of your work in R. That will complement your writing by illustrating your logic in the most fundamental way. Feel free to comment your code using the # symbol if you think that it will help me to follow your thinking. The focus should be on the functional R code required to do the modelling project. Inevitably you will have tried many things, not all of which will have worked. You do not need to include the false starts and coding mistakes in what you submit. I am just interested in seeing the code that works. ENST90044—Monitoring Ecosystems and Their Values 3 Label your figures: This includes axis labels, figure numbers, and figure captions. This will make it easier to follow and understand your figures throughout the text. Do not use higher-order polynomials: Because they are always a bad idea!