Hugging Face's logo

Datasets:
community-datasets
/
so_stacksample

Tasks:
Text2Text Generation
Sub-tasks:
abstractive-qa
open-domain-abstractive-qa
Languages:
English
Size:
1M<n<10M
License:
Dataset card Files Community
1
so_stacksample / README.md
lhoestq's picture
lhoestq HF staff
add dataset_info in dataset metadata
dfdf39f
preview code
|
raw
history blame contribute delete
8.24 kB
metadata
annotations_creators:
  - no-annotation
language_creators:
  - crowdsourced
language:
  - en
license:
  - cc-by-sa-3.0
multilinguality:
  - monolingual
size_categories:
  - 1M<n<10M
source_datasets:
  - original
task_categories:
  - text2text-generation
task_ids:
  - abstractive-qa
  - open-domain-abstractive-qa
paperswithcode_id: null
pretty_name: SO StackSample
dataset_info:
  - config_name: Answers
    features:
      - name: Id
        dtype: int32
      - name: OwnerUserId
        dtype: int32
      - name: CreationDate
        dtype: string
      - name: ParentId
        dtype: int32
      - name: Score
        dtype: int32
      - name: Body
        dtype: string
    splits:
      - name: Answers
        num_bytes: 1583232304
        num_examples: 2014516
    download_size: 0
    dataset_size: 1583232304
  - config_name: Questions
    features:
      - name: Id
        dtype: int32
      - name: OwnerUserId
        dtype: int32
      - name: CreationDate
        dtype: string
      - name: ClosedDate
        dtype: string
      - name: Score
        dtype: int32
      - name: Title
        dtype: string
      - name: Body
        dtype: string
    splits:
      - name: Questions
        num_bytes: 1913896893
        num_examples: 1264216
    download_size: 0
    dataset_size: 1913896893
  - config_name: Tags
    features:
      - name: Id
        dtype: int32
      - name: Tag
        dtype: string
    splits:
      - name: Tags
        num_bytes: 58816824
        num_examples: 3750994
    download_size: 0
    dataset_size: 58816824

Dataset Card for SO StackSample

Table of Contents

Dataset Description

Dataset Summary

Dataset with the text of 10% of questions and answers from the Stack Overflow programming Q&A website.

This is organized as three tables:

Questions table contains the title, body, creation date, closed date (if applicable), score, and owner ID for all non-deleted Stack Overflow questions whose Id is a multiple of 10. Answers table contains the body, creation date, score, and owner ID for each of the answers to these questions. The ParentId column links back to the Questions table. Tags table contains the tags on each of these questions.

Supported Tasks and Leaderboards

Example projects include:

  • Identifying tags from question text
  • Predicting whether questions will be upvoted, downvoted, or closed based on their text
  • Predicting how long questions will take to answer
  • Open Domain Q/A

Languages

English (en) and Programming Languages.

Dataset Structure

Data Instances

For Answers:

{
  "Id": { # Unique ID given to the Answer post
    "feature_type": "Value",
    "dtype": "int32"
  },
  "OwnerUserId": { # The UserID of the person who generated the Answer on StackOverflow. -1 means NA
    "feature_type": "Value",
    "dtype": "int32"
  },
  "CreationDate": { # The date the Answer was generated. Follows standard datetime format.
    "feature_type": "Value",
    "dtype": "string"
  },
  "ParentId": { # Refers to the `Id` of the Question the Answer belong to.
    "feature_type": "Value",
    "dtype": "int32"
  },
  "Score": { # The sum of up and down votes given to the Answer. Can be negative.
    "feature_type": "Value",
    "dtype": "int32"
  },
  "Body": { # The body content of the Answer.
    "feature_type": "Value",
    "dtype": "string"
  }
}

For Questions:

{
  "Id": { # Unique ID given to the Question post
    "feature_type": "Value",
    "dtype": "int32"
  },
  "OwnerUserId": { # The UserID of the person who generated the Question on StackOverflow. -1 means NA.
    "feature_type": "Value",
    "dtype": "int32"
  },
  "CreationDate": { # The date the Question was generated. Follows standard datetime format.
    "feature_type": "Value",
    "dtype": "string"
  },
  "ClosedDate": { # The date the Question was generated. Follows standard datetime format. Can be NA.
    "feature_type": "Value",
    "dtype": "string"
  },
  "Score": { # The sum of up and down votes given to the Question. Can be negative.
    "feature_type": "Value",
    "dtype": "int32"
  },
  "Title": { # The title of the Question.
    "feature_type": "Value",
    "dtype": "string"
  },
  "Body": { # The body content of the Question.
    "feature_type": "Value",
    "dtype": "string"
  }
}

For Tags:

{
  "Id": { # ID of the Question the tag belongs to
    "feature_type": "Value",
    "dtype": "int32"
  },
  "Tag": { # The tag name
    "feature_type": "Value",
    "dtype": "string"
  }
}

`

Data Fields

For Answers: -Id: Unique ID given to the Answer post OwnerUserId: The UserID of the person who generated the Answer on StackOverflow. -1 means NA "CreationDate": The date the Answer was generated. Follows standard datetime format. "ParentId": Refers to the Id of the Question the Answer belong to. "Score": The sum of up and down votes given to the Answer. Can be negative. "Body": The body content of the Answer.

For Questions:

  • Id: Unique ID given to the Question post.
  • OwnerUserId: The UserID of the person who generated the Question on StackOverflow. -1 means NA.
  • CreationDate: The date the Question was generated. Follows standard datetime format.
  • ClosedDate: The date the Question was generated. Follows standard datetime format. Can be NA.
  • Score: The sum of up and down votes given to the Question. Can be negative.
  • Title: {The title of the Question.
  • Body: The body content of the Question.

For Tags:

  • Id: ID of the Question the tag belongs to.
  • Tag: The tag name.

Data Splits

The dataset has 3 splits:

  • Answers
  • Questions
  • Tags

Dataset Creation

Curation Rationale

Datasets of all R questions and all Python questions are also available on Kaggle, but this dataset is especially useful for analyses that span many languages.

Source Data

Initial Data Collection and Normalization

[More Information Needed]

Who are the source language producers?

StackOverflow Users.

Annotations

Annotation process

[More Information Needed]

Who are the annotators?

[More Information Needed]

Personal and Sensitive Information

This data contains information that can identify individual users of StackOverflow. The information is self-reported.

[Needs More Information]

Considerations for Using the Data

Social Impact of Dataset

StackOverflow answers are not guaranteed to be safe, secure, or correct. Some answers may purposefully be insecure as is done in this https://stackoverflow.com/a/35571883/5768407 answer from user zys, where they show a solution to purposefully bypass Google Play store security checks. Such answers can lead to biased models that use this data and can further propogate unsafe and insecure programming practices.

[Needs More Information]

Discussion of Biases

[Needs More Information]

Other Known Limitations

[Needs More Information]

Additional Information

Dataset Curators

[Needs More Information]

Licensing Information

All Stack Overflow user contributions are licensed under CC-BY-SA 3.0 with attribution required.

Citation Information

The content is from Stack Overflow.

Contributions

Thanks to @ncoop57 for adding this dataset.