add dataset readme

code/myriad/loam_paper/dataset/README.md

@@ -0,0 +1,34 @@
+# Dataset
+
+This is the **Myriad People** dataset. All these files have been generated by running `script.py` in the `../mining` folder. Here are their descriptions.
+
+`all_loggedin_contributors.json`: list of all logged-in (i.e. not anonymous) GitHub contributors, with:
+- `type`: type of contributor, `User` or `Bot`
+- `id`: GitHub username
+- `contributions`: list of repositories they contributed to, with:
+    - `repo_name`: name of repository
+    - `contributions`: number of contributions that they made to this project
+
+`categories_info.json`: list of categories, with:
+- `category`: name of the category
+- `repos`: list of names of repositories in that category
+
+`repos_info.json`: list of all repositories for which the GitHub API managed to fetch the data, with:
+- `name`: name of the repository
+- `category`:  category it belongs to
+- `exclusivity`: either the name of an artwork if this repository was exclusively used in that artwork, or `null` if it was used in at least two artworks
+- `created_at`: creation date of the repository, in the Python `datetime` format
+- `total_contributions`: total number of contributions
+- `anonymous_contributors`: number of anonymous contributors
+- `loggedin_contributors`: number of logged-in contributors
+
+`gh_api_failures.json`: list of repositories for which the GitHub API failed (because they are too big), with `name`, `category` and `exclusivity`, as in `repos_info.json`
+
+`individual_repos` folder: one file per repository, in the format `owner&name.json`, with:
+- `repo_name`: name of the repository
+- `contributors`: list of contributors, with:
+    - `type`: type of contributor, `User` or `Bot`
+    - `id`: GitHub username
+    - `contributions`: number of contributions that they made to this project
+
+In all the files, repository names attributes are in the format `owner/name`.