Geographers study big things, cities, forests, snowpack, climate. The systems geographers study are usually too big to directly observe. Geographers rely on data from samples, sensors, surveys, and satellites. Statistics is the art and science of building knowledge from data. This class teaches you how to work with data and how to use it to gain insight into geographic systems. Often geographic phenomena are characterized by complex interactions accross space, time, and scale. This is a second course in statistics and it assumes some knowledge of basic statistical concepts.
Geography is a diverse discipline but in spite this diversity there are set of core analytical techniques that are used across the discipline. This class will focus on these core analytical methods. We will, inevitably, cover subject matter that is not specific to your current area of study and we may miss some things that are central to your sub-field. However, the broad exposure to analytical methods provided by this class will (hopefully) provide a platform for you to learn discipline specific methods in the future.
In this class we focus on developing an intuitive understanding of statistical methods, generally we do this via discussion and example, not formal proofs. On the one hand, I do not want statistical methods to be "black-boxes", on the other hand, I don't want to spend weeks of class time on mathemetical detals. Therefore, this course emphasizes applied analysis through hands-on experiences that aim to provide practical understanding of methods. You should leave this course with confidence in the methods we have discussed and an appreciation for how these statistical methods are applied in "real" geographic research. I will emphasize conceptual understanding, how to implement statistical tests in R and interpret the output.
This is a course for geographers. You do not have to be a geography student to benefit from this class. However, this class assumes an interest in spatial problems and spatial data. Many of the techniques we will learn are not explicitly spatial, however we will, whenever possible, discuss the application of these techniques to spatial problems.
By the end of the course, I want you to know how to select the appropriate statistical method to answer a research question, analyze data, and correctly interpret and write-up the results of your analysis. The course objectives are:
To develop "statistical literacy," a working understanding of statistics that can help in critically evaluating data-driven results in the discipline of geography (or ecology, etc...).
To obtain a rich set of statistical tools for data analysis, with an understanding of the how to choose appropriate tools and implement them in statistical software.
To enable you to confidently and carefully interpret the results of data analyses and clearly communicate those results.
To provide practical experience in using real sets of data addressing meaningful research questions.
This class depends heavily on the R programming language, you can expect to spend several hours each week programming in R.
I am here to organize the course and introduce you to the topics and readings we will examine. I don't have all the answers and I don't pretend to have all the answers, but I will share with you what I know. I will do my best to make the course interesting, relevant, and challenging. That being said, it's important that you understand that you have the most important role in making GEOG 5023 a success. You will determine how much you actually get out of this course. Doing the readings outlined, and coming to class and labs ready to think and participate in group discussions puts you in the best position to benefit from what this course offers. I encourage you to make full use of the learning opportunities that this class presents I am very open to feedback and I would like to help you overcome any problems. If there is ever anything you don't understand please get in touch. If you'd like more detail on something, just ask.
I request that you post all questions related to lectures, labs, exams, and the final project on the class Piazza site . The reason for this is that it allows others to benefit from your question, allows students to answer, and prevents duplication of questions. Petra and I will post responses to Piazza questions. If you have a question or concern that you'd like to keep private please don't hesitate to contact the instructors directly.
Students enrolled in this course must have completed an introductory statistics course (e.g. GEOG 3023, APPM 4570, ECON 3818, PSYC 3101, SOCY 4061, EDUC 5716). This course satisfies the requirement for quantitative methods for MA and PhD students in Geography.
We will be using quite a few books this semester making purchasing them all is an expernsive proposition. Assigned readings will be available as reserves. I think, R for Everyone: Advanced Analytics and Graphics, a small book might be a useful desk reference, especially for those new to programming in R. Lab materials are available here.
5% of Total Course Grade (GRADS), 15% for Undergrads
Labs will vary in scope. Some will take one week (short labs) to complete, others will take two weeks (long labs). Generally, short lab assignments provide more direction than the two week long labs. Short labs will generally include code to get you started, long labs will not. For both short and long labs I will provide you a dataset, a set of questions, and some programming advice. To complete the lab you will have to figure out how to use the statistical techniques and the software we've covered in class to answer a set of broad questions. Labs are open ended and designed to allow you significant latitude and room for creativity.
Lab write-ups are expected to look very similar to a journal article's results and discussion section. I will pass out a grading rubric and an example of a good lab before the first lab so you understand how labs will be graded. You may work on your lab in groups but you should submit individual assignments. When a lab has been completed collaboratively identify your collaborators and how they contributed to the final product. Feel free to use any online or offline resources that you find helpful. Feel free to share ideas and/or ask questions on the Piazza website. The only way you can cheat on a lab is taking a classmates code/ideas without their consent. Anything online is fair game.
Article Commentaries and Disscussant
Grad students are expected to write and present three article commentaries. You may select articles yourself. Articles should employ a methodological approch covered in the class or one that you feel comfortable explaining to the class. Your commentaries should address the questions in the "reading scientifc articles worksheet." You must prepare a brief summary of the article and a 5-10 minute presentation. Each commentary will be assigned a discussant. The discussant role is to ask critical questions about your commentary and to kick-off class discussion. You must give the discussant your commentary in the class session before it is presented. Commentaries and discussions should be a critical evaluation of the research methods used in the article. The idea here is to have discussions that provide insight into how methods are used in the scientifc literature.
There will be two exams. Exams will focus on the interepretation of model output.
Historically, this class has included an individual final project. In spring 2015 we'll be trying a new approach, inspired by kaggle.com. We will, via some democratic mechanism, select a class project. The class will be divided into teams, each team will complete the project to the best of their ability. Final projects will be graded as two labs; One lab will be for the analysis (code and methods), one for the presentation of results.
Late assignments (labs, commentaries) up to 1 week late will be downgraded 20%, 100% thereafter. Students must complete all lab assignments to receive a passing grade, even if they are submitted too late to receive any points. Exams must be entirely your own work.