GLARE: Google Apps Arabic Reviews

Dataset and Code of "GLARE: Google Apps Arabic Reviews" paper.

You can download the paper via: [Github]

Paper Summary

We introduce GLARE: Google Apps Arabic Reviews dataset. A collection of 76M reviews from 9,980 Android apps collected from Google PlayStore Saudi store.

Preparation

Below is details about each file, please ensure that you have enough storage before downloading the data.

Data Type	File Name	File Size	File Type
raw	apps	4.1 MB	CSV
raw	reviews	17 GB	CSV
raw	categories/	4.3 MB	CSV
engineered	apps	3.8 MB	CSV
engineered	reviews	21.9 GB	CSV
engineered	vocabulary	530.5 MB	CSV

File Specifications

• apps.csv: File that contains apps metadata.
• reviews.csv: File that contains reviews and reviews metadata.
• categories/: Folder that contains 59 CSV files, each file corresponds to one category with apps and apps metadata scrapped from top 200 free apps for that category.
• vocabulary.csv: File that contains vocabulary set generated from reviews with additional engineered features (word length, word frequency, has noise or digits, ..etc.)

Raw Data

Apps Metadata

{
    "title":"application name/title",
    "app_id":"application unique identifier",
    "url":"application url at Google PlayStore",
    "icon":"url for image object",
    "developer":"developer name",
    "developer_id":"developer unique identifier",
    "summary":"short description of the application",
    "rating":"application accumulated rating"
 }

Reviews Metadata

{
   "at":"review datetime",
   "content":"review text",
   "replied_at":"developer reply datetime",
   "reply_content":"developer reply content",
   "review_created_version":"user application version during the time of review",
   "review_id":"review unique identifier",
   "rating":"user rating",
   "thumbs_up_count":"number of users that agree with the reviewer",
   "user_name":"user display name",
   "app_id":"application unique identifier"
}

Engineered Data

Apps Metadata

Same as apps.csv in raw data with the following additions:

{
   "reviews_count":"number of reviews for the application",
   "categories":"list of application categories",
   "categories_count":"number of application categories"

}

Reviews Metadata

Same as reviews.csv in raw data with the following additions:

{
  "tokenized_review":"list of review words tokenized on white-space",
  "words_count":"number of words in review"
}

Vocabulary

{
   "word":"term text",
   "length":"word characters count",
   "frequency":"word occurrences in the reviews dataset",
   "has_noise":"true or false if word contains anything non-arabic alphanumeric",
   "noise":"list of noise (anything non-arabic alphanumeric) in the word",
   "has_digits":"true or false if word contains arabic or hindi digits",
   "digits":"list of digits in the word"
}

Folders Structure

• Data are prepared as raw data or engineered data.
• Download the dataset files: Google Drive | Zenodo | Alternative Google Drive
• The directory structure is as follow:

data
└── raw
   ├── apps.csv
   ├── reviews.csv
   └── categories/
└── engineered
   ├── apps.csv
   ├── reviews.csv
   └── vocabulary.csv

Citation

If you use this dataset please cite as:

@dataset{alghamdi_fatima_2022_6457824,
  author       = {AlGhamdi, Fatima and
                  Mohammed, Reem and
                  Al-Khalifa, Hend and
                  Alowisheq, Areeb},
  title        = {GLARE: Google Apps Arabic Reviews Dataset},
  month        = apr,
  year         = 2022,
  publisher    = {Zenodo},
  version      = {1.0},
  doi          = {10.5281/zenodo.6457824},
  url          = {https://doi.org/10.5281/zenodo.6457824}
}

License

This work is licensed under a Creative Commons Attribution 4.0 International License.