Project

For the project in this class, you will analyze data from a Norwegian study on air pollution, conducted at Alnabru in Oslo, Norway, from October 2001 to August 2003, and write a paper on your conclusions. The data, along with a detailed description of the variables, what they mean, and how to read the data into R are available here.

The objective of this project is to study the relationship between the logarithm of the concentration of nitrogen dioxide, NO₂ (particles), and several traffic volume and meteorological variables. You should identify a statistical model that accurately describes the log NO₂ concentration, with the explanatory variables explaining as much variability as possible, but without making it so complex that you overfit (make it unusable for other data collected through the same mechanism). You are not required to use all the variables in your final model, and you should be mindful of the relationships among some of the variables.

There are two final delivarables for the project:

1. A paper describing your problem, the analysis you conducted, and your conclusions
2. Supplementary materials with R output and diagnostic plots for your model

The report and supplementary materials should both be generated by an R markdown document, and I will ask you to submit both the R markdown document and the resulting knitted pdf.

Deadlines

• 11:59 PM EDT Mon., April 20: At this check point, you should have fit a multiple regression model, checked residual diagnostics, and taken steps to address any evident problems. Submit by uploading to google drive (Spring 2020 Projects) and sending me an email letting me know you have done that.
• 11:59 PM EDT Fri., April 24: Final submission of R markdown file and pdf for report and supplementary materials. Submit by uploading to google drive (Spring 2020 Projects) and sending me an email letting me know you have done that.

Guidelines for the final report

• Overall, the project report should be written in clear, concise prose. No R code should be shown (To hide R code in the knitted Markdown file, the first part of each r chunk should be: {r echo=FALSE})
• We will use a structure that is similar to a standard scientific report, though your write up will likely be somewhat shorter than a typical journal article. Please follow the structure below:
1. Title
2. Summary: an introduction to the problem we are addressing, a brief description of the methods you consider, and a summary of the results. Aim for 1 paragraph.
3. Data: a brief summary of key features of the dataset. You should define each variable that will be used (to the level that it is possible to do this, given the information provided about the data). Also include a few plots showing a few key insights about the data set. Note that there will probably not be enough space to present every plot you make during the course of conducting your analysis; you will have to select a small number of the most informative plots to include. These plots should be briefly discussed in the text. At least a few sentences of context and description of the dataset should be included (how were the data collected? What was measured?), and the number of observations in the data set should be stated. Aim for about 1-2 pages. There should be enough detail that the scope of conclusions from your analysis can be assessed.
4. Methods: a description of the statistical model used in your analysis. Describe any transformations or other special things you had to do. Aim for a page or less.
5. Results: a presentation of your results. This should include a paragraph or two stating the results of the analysis with minimal interpretation. Aim for less than a page.
6. Discussion: summarize your work, its limitations, and possible future steps/improvements. Address the answers to the problem you outlined in your summary and the scope of your conclusions. This can be a page or two.
7. References: cite all sources in a standard format.

Items 1-6 above will probably require between 5 and 10 pages, including figures and tables. Please do not go over 10 pages. If your report is looking like it will be less than 5 pages please run it by me and make sure you’re discussing everything in enough detail. You should not change the font size or margins from the defaults for R markdown documents.

Grading and Assessment Criteria

The project grade makes up 10% of the final grade for the class. Here are some things I’ll be considering:

1. Technical Mastery: Do you demonstrate that you understand the methods you are using? Does the submitted R code work correctly? Can I knit the submitted R markdown files to generate the submitted pdf file?
2. Writing: How effectively does the written report communicate the goals, procedures, and results of the study? Are the claims adequately discussed and supported? How well is the report structured and organized (this should not be a problem if you follow the structure I laid out!)? Are all of the figures and tables numbered and appropriately referenced? Does the writing style enhance what you are trying to communicate? How well is the report edited?
3. Statisical Analysis: Are the chosen analyses appropriate for the variables/relationships under investigation, and are the assumptions underlying these analyses met? Do the analyses involve fitting and interpreting a multiple regression model? Are the analyses carried out correctly? Was the appropriateness of the model assessed using diagnostic plots? Is there an effective mix of graphical, numerical, and inferential analyses?
4. Conclusions: Are the stated conclusions supported and justified by the analysis? Can the effects of confounding variables be controlled for (if not, is that discussed as a limitation of the analysis)? Is the scope of conclusions properly addressed?