Project Website hosted on GitHub Pages [15 points]

You will create an index.md file to serve as the landing page for the report, and link to report chapters from that document. Individual chapters should be stored as separate RMD files (rendered as HTML files for the report). You can link to rendered HTML files by referencing them in the proper sub-directories:

• http://site-url.com (will load the index file)
• http://site-url/analysis/file-name.html (file in analysis/ sub-directory)
• Website is active and live
  • All links work properly
• GitHub pages template is clean and effective
  • Custom CSS used for consistent report style
• Landing page includes:
  • a table of contents
  • a link to files on the GitHub repo with description of content
  • replication instructions (software needed, how to access files, etc)
  • license info
• About Us page

Data [10 points]

• Original data sources are documented and raw data used in the project has been archived
• Data steps clearly walk through the process of creating new variables, cleaning data, and joining tables.
  • These steps should exist in a standalone .R script that hosts all of these steps and produces .rds files that store data used throughout your project.
  • This allows you to avoid copying and pasting the same cleaning code while holding one file as the source of truth for your data cleaning
• README.md in the data folder that explains the organization of the files

Analysis [15 points]

• Quality of models and presentation of the steps to produce results:
  • Chapter on descriptive analysis of neighborhood change
  • Chapter predicting change with neighborhood characteristics
  • Chapter describing tax credit programs (NMTC & LIHTC)
  • Predictive models after adding tax credit programs
• Data manifest:
  • A table that reports the initial sample size in the raw data set, the sample size in the final model, and explain where and why each observation was dropped
• All results are easily reproducible
  • Packages and versions reported courtesy of the renv package.
  • Clear description of which data is used for each model
    • Appropriate references to detailed data step files when helpful
  • Functions and arguments are explained when they are not clear
    • Specialized packages or functions
    • Custom functions

Portability [10 points]

• Can anyone run project code on their computer without having to change working directories or other settings?
• Note that Windows and Mac use different styles for paths (forward vs backwards slash) so functions like here::here() can help avoid those issues.
• Report the R environment and package versions used to create the analysis
  • Usage of renv package satisifes this requirement

Uniform directory structure

Every team’s directory structure should look like this:

. (root/main/project directory, such as /Users/pacman/asu/cpp-528-spr-2021)
├── README.md
├── analysis
│   └── README.md
├── assets
│   ├── README.md
│   ├── css
│   │   └── README.md
│   └── images
│       └── README.md
├── <RStudio Project Label>.Rproj
├── data
│   ├── README.md
│   ├── raw
│   │   ├── LTDB-DATA-DICTIONARY.xlsx
│   │   ├── LTDB-codebook.pdf
│   │   ├── LTDB_County_1980_Global_Neighborhood.csv
│   │   ├── LTDB_County_1990_Global_Neighborhood.csv
│   │   ├── LTDB_County_2000_Global_Neighborhood.csv
│   │   ├── LTDB_County_2010_Global_Neighborhood.csv
│   │   ├── LTDB_Std_1970_fullcount.csv
│   │   ├── LTDB_Std_1980_fullcount.csv
│   │   ├── LTDB_Std_1990_fullcount.csv
│   │   ├── LTDB_Std_2000_fullcount.csv
│   │   ├── LTDB_Std_2010_fullcount.csv
│   │   ├── LTDB_Std_All_Sample.zip
│   │   ├── LTDB_Std_All_fullcount.zip
│   │   ├── README.md
│   │   ├── cbsa-crosswalk.rds
│   │   ├── ltdb_data_dictionary.csv
│   │   ├── ltdb_std_1970_sample.csv
│   │   ├── ltdb_std_1980_sample.csv
│   │   ├── ltdb_std_1990_sample.csv
│   │   ├── ltdb_std_2000_sample.csv
│   │   └── ltdb_std_2010_sample.csv
│   ├── rodeo
│   │   ├── LTDB-1970.rds
│   │   ├── LTDB-1980.rds
│   │   ├── LTDB-1990.rds
│   │   ├── LTDB-2000.rds
│   │   ├── LTDB-2010.rds
│   │   ├── LTDB-META-DATA.rds
│   │   └── README.md
│   └── wrangling
│       └── README.md
├── labs
│   ├── wk01
│       └── README.md
│   ├── wk02
│       └── README.md
│   ├── wk03
│       └── README.md
│   ├── wk04
│       └── README.md
│   ├── wk05
│       └── README.md
│   ├── wk06
│       └── README.md
│   └── wk07
│       └── README.md

When creating files/folders, here are some helpful guidelines:

• Never capitalize unless necessary or helpful for emphasis
• Use an underscore (_) instead of a space in file and folder names since spaces can cause a lot of problems in paths and are usually replaced with arbitrary characters.
• Try to have short but meaningful file names that are memorable
• Add a README.md file inside of each folder with, at minimum, a one-sentence description of the directory.
• If files need to run in a specific order, or they represent things like chapters of a book, consider naming them something like step_01, step_02, etc.
• File names:
  • Consistent naming throughout project, including rules for capitalization and dates
  • When order matters (for example steps in analysis or chapters of a report) the file order matches order in which they should be run
  • Effective use of leading zeros to maintain proper file order (09, 10, 11, … )
• Clear system to keep track of the current version and archive revisions

It is helpful to have only the current version of a file in the project directory, and move all old versions to an old-files/ or archive/ folder if you need to reference them. If you have multiple versions of the same file in your project folder there is a high probability that one team member accidentally uses the wrong version (especially if they have not synced their project directory). It also prevents you from reading old versions of the file into a script.

README Files [10 points]

Main README.md in the repository

• Overview of the project and the code
• Links to appropriate pages in the report
• Be kind and include convenient code to install all packages you use in the report
  • Since you’re using renv package, you should have a renv.lock file that stores your packages and their dependencies. Therefore, just make sure you explictly state that the reader has to use renv::restore() to download the specific versions of the packages used in your project.
  • Only one line of code but it is powerful!
• If you write a custom package, include the command to install the package from GitHub so it can be easily copied and pasted
• Contact information for people that will be maintaining the project. At most, ASU emails; at minimum, links to your personal GitHub pages.
• If you include a shiny app, include the launch from GitHub command in the README file so that it is accessible
• Sub-folder README files
  • Should explain the purpose and organization of the directory
  • Include instructions for using the files in the folder

Readable Code [5 points]

• Everyone follows the same code style
  • Clean and consistent feel
  • Indents used to group code appropriately
• Objects follow consistent naming conventions
  • Data sets are nouns, functions are verbs
• Code is commented appropriately throughout the files

Citations [5 points]

• All sources are appropriately cited
• Bibtex citation database is created for the project
• R Markdown generates a bibliography automatically
  • Include a citation list at the end of each chapter License [5 points]
• Select an appropriate open-source license for the project and include it in the repository.
• Add a LICENSE file on GitHub and include the appropriate licensing info

Citation Integration in RMD

Here is an example .rmd file that does two things:

1. It imports a .bib file inside the YAML header (see line 6); and
2. It prints out all the citations at the end of the .rmd file (see line 35).

For context, checkout RStudio’s resource guide.