Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Dataset description #1

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Dataset description #1

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
130 changes: 129 additions & 1 deletion recsys-2017/index.html
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -179,7 +179,135 @@ <h2>Evaluation Metrics</h2>

<div id="dataset" class="lead">
<h2>Dataset</h2>
<div class="alert alert-success" role="alert">Coming soon...</div>
<p>
In the challenge, the task of the participants will be the following: given a XING user, the recommender should predict those job postings (items) that the user will interact with in the next week.
The traing dataset is supposed to be used for experimenting and training your models. You can split the
interaction data into training and test data. For example: you can leave out the last
complete week (of the year) from the interaction data and then try to predict whether
a given user will _positively_ interact with an item within that week. Relevant items
are those items on which a user clicked, bookmarked or replied (`interaction_type` = 1, 2 or 3).
</p>
<h3> Anonymization, pseudonymization, noise </h3>
<p>
The traing dataset is a semi-synthetic sample of XING's dataset, i.e. it is not complete and enriched with
noise in order anonymize the data. For example:
<ul>
<li> the dataset contains artifical users </li>
<li> the dataset contains only a fraction of XING users and job postings </li>
<li> IDs are used instead of raw text for almost all attribute values (pseudonymization) </li>
<li> some attributes of the users may have been removed or flipped to <i>NULL / unknown</i>. </li>
<li> not all interactions of a user are contained in the dataset </li>
<li> some of the interactions are artificial (= have actually not been performed by the user) </li>
<li> timestamps have been shifted (but the order of interactions is kept) </li>
</ul>
Attempting to identify users or to reveal any private information about the users or information about
the business from which the data is coming from is strictly forbidden (cf. [Rules](http://2016.recsyschallenge.com/)).
</p>

<h3> Interactions </h3>
<p>
Interactions are all transactions between a user and an item including
recruiter interest as well as impressions.
Fields:
<ul>
<li> `user_id` ID of the user who performed the interaction (points to `users.id`) </li>
<li> `item_id` ID of the item on which the interaction was performed (points to `items.id`) </li>
<li> `created_at` a unix time stamp timestamp representing the time when the interaction got created </li>
<li> `interaction_type` the type of interaction that was performed on the item: <ul>
<li> 0 = <b>XING showed this item to a user (= impression)</b> </li>
<li> 1 = the user clicked on the item </li>
<li> 2 = the user bookmarked the item on XING </li>
<li> 3 = the user clicked on the _reply button_ or _application form button_ that is shown on some job postings</li>
<li> 4 = the user deleted a recommendation from his/her list of recommendation (clicking on "x") which has the effect that the recommendation will no longer been shown to the user and that a new recommendation item will be loaded and displayed to the user </li>
<li> 5 = <b>a recruiter from the items company showed interest into the user.</b> (e.g. clicked on the profile)</li>
</ul></li>
</ul>
</p>

<h3> Users </h3>
<p>
Details about those users who appear in the above datasets. Fields:
<ul>
<li> `id` anonymized ID of the user (referenced as `user_id` in the other datasets above) </li>
<li> `jobroles` comma-separated list of jobrole terms (numeric IDs) that were extracted from the user's current job titles </li>
<li> `career_level` career level ID (e.g. beginner, experienced, manager): <ul>
<li> 0 = unknown </li>
<li> 1 = Student/Intern </li>
<li> 2 = Entry Level (Beginner) </li>
<li> 3 = Professional/Experienced </li>
<li> 4 = Manager (Manager/Supervisor) </li>
<li> 5 = Executive (VP, SVP, etc.) </li>
<li> 6 = Senior Executive (CEO, CFO, President)</li></ul></li>
<li> `discipline_id` anonymized IDs represent disciplines such as "Consulting", "HR", etc. </li>
<li> `industry_id` anonymized IDs represent industries such as "Internet", "Automotive", "Finance", etc. </li>
<li> `country` describes the country in which the user is currently working<ul>
<li> `de` = Germany </li>
<li> `at` = Austria</li>
<li> `ch` = Switzerland </li>
<li> `non dach` = non of the above countries</li></ul></li>
<li> `region` is specified for some users who have as country `de`. Meaning of the regions see below</li>
<li> `experience_n_entries_class` identifies the number of CV entries that the user has listed as _work experiences_ <ul>
<li> 0 = no entries </li>
<li> 1 = 1<li>2 entries </li>
<li> 2 = 3<li>4 entries </li>
<li> 3 = 5 or more entries</li></ul></li>
<li> `experience_years_experience` is the estimated number of years of work experience that the user has<ul>
<li> 0 = unknown </li>
<li> 1 = less than 1 year </li>
<li> 2 = 1 - 3 years </li>
<li> 3 = 3 - 5 years </li>
<li> 4 = 5 - 10 years </li>
<li> 5 = 10 - 20 years </li>
<li> 6 = more than 20 years </li> </ul> </li>
<li> `experience_years_in_current` is the estimated number of years that the user is already working in her current job. Meaning of numbers: same as `experience_years_experience` </li>
<li> `edu_degree` estimated university degree of the user<ul>
<li> 0 or NULL = unknown </li>
<li> 1 = bachelor </li>
<li> 2 = master </li>
<li> 3 = phd </li> </ul> </li>
<li> `edu_fieldofstudies` comma<li>separated fields of studies that the user studied. `0` means "unknown" and `edu_fieldofstudies > 0` entries refer to broad field of studies such as _Engineering_, _Economics and Legal_, ... </li>
<li> <b>`wtcj` predicted willingness to change jobs</b> <ul>
<li> 0 XING predicts the user won't change jobs soon </li>
<li> 1 XING predicts the user is interested in changing his current position </li> </ul> </li>
<li> <b>`premium` the user subscribed to XING's payed premium membership</b> <ul>
<li> 0 no subscription </li>
<li> 1 active subscription </li> </ul> </li>
</ul>
</p>

<h3> Items </h3>
<p>
Details about the job postings that were and should be recommended to the users.
<ul>
<li> `id` anonymized ID of the item (referenced as `item_id` in the other datasets above) </li>
<li> `industry_id` anonymized IDs represent industries such as "Internet", "Automotive", "Finance", etc.</li>
<li> `discipline_id` anonymized IDs represent disciplines such as "Consulting", "HR", etc. </li>
<li> <b>`is_paid` indicates that the posting is a paid for by a compnay </b> </li>
<li> `career_level` career level ID (e.g. beginner, experienced, manager) <ul>
<li> 0 = unknown </li>
<li> 1 = Student/Intern </li>
<li> 2 = Entry Level (Beginner) </li>
<li> 3 = Professional/Experienced </li>
<li> 4 = Manager (Manager/Supervisor)</li>
<li> 5 = Executive (VP, SVP, etc.) </li>
<li> 6 = Senior Executive (CEO, CFO, President) </li> </ul> </li>
<li> `country` code of the country in which the job is offered </li>
<li> `latitude` latitude information (rounded to ca. 10km) </li>
<li> `longitude` longitude information (rounded to ca. 10km) </li>
<li> `region` is specified for some users who have as country `de`. Meaning of the regions: see below. </li>
<li> `employment` the type of emploment <ul>
<li> 0 = unknown </li>
<li> 1 = full-time </li>
<li> 2 = part-time </li>
<li> 3 = freelancer </li>
<li> 4 = intern </li>
<li> 5 = voluntary </li>
<li> `created_at` a unix time stamp timestamp representing the time when the interaction got created </li>
<li> `title` concepts that have been extracted from the job title of the job posting (numeric IDs) </li>
<li> `tags` concepts that have been extracted from the tags, skills or company name </li>
</ul> </li>
</ul>
</p>
</div>


Expand Down