Submission

In a nut shell...

1. Fork this repository.
2. Open a pull request from your fork to the submission branch.
3. As you complete tasks, commit your results.

Materials

There are three folders in the repository that have been pre-created for you.

AITP2019-DS_Challenge/
├── README.md
├── code/
├── docs/
└── outputs/

3 directories, 1 file

At the end of the day, you will be tasked with presenting the work that you've done to the judging personnel. Further, all of the materials you have produced will be used for assessing your work. The three folders provided give a consistent, and largely self explanatory, framework for organizing your submission.

Any code that is used should be organized within that folder. The outputs generated by that code should be organized below that folder.

Note! Do not commit large data files to Git, those should stay in your Scratch space on the cluster. See Resources.

You should develop the story telling aspect of your work in markdown files within the docs folder. It is expected that you will make use of these during your presentation. (No slide shows please). The advantage of markdown is its simplicity (you don't have time to be polishing a slide deck) and they can be rendered nicely on the web by Github. This helps the judging personnel review your work without having to download or run any software.

Within the docs folder, please organize your markdown pages by the Stages outlined on the Challenge wiki page. If you pursue multiple lines of inquery, isolate those as subfolders within the docs folder.

Example:

AITP2019-DS_Challenge/
├── README.md
├── code/...
├── docs/
   ├── indiana/
      ├── 01_extract.md
      ├── 02_validation.md
      ├── 03_analysis.md
      └── 04.visualizations.md
   └── something_else/
      ├── 01_extract.md
      ├── 02_validation.md
      ├── 03_analysis.md
      └── 04.visualizations.md
└── outputs/...

Note: If you use Jupyter notebooks to do your work, these can be rendered by Github and it is acceptible to organize your submission entirely within the docs folder (with code and outputs inlined with your prose in the notebook.

Using Github

Version control software, such as Git, has become an integral part of modern data science. It offers a mechanism to allow for multi-person teams to work on the same code and documentation on different computers at the same time. If you are unfamiliar with how to use Git and Github, a brief introduction at the beginning of the competition should be sufficient, as well as whatever help is necessary during the day.

For multi-person teams, a single person (doesn't matter who) should fork this repository and open the pull request. Then, add the other members of the team as Collaborators (under settings). This way, they can push changes to the same repository.

The git command is available on the Brown cluster and it is recommended that you login there to create your "local" clone.

---

Previous: Challenge   |   Next: Scoring