Mansour-Essgaer/Libya-Telecom-companies-corpus

Libyana, AlMadar Aljadid, and Libya Telecom and Technology (LTT) corpus data which were collected from Twitter during the Covid 19 outbreak # Libyan Telecom Customer Satisfaction Corpus ### A Sentiment Analysis Dataset for Libyan Telecommunication Companies ## 📌 Overview The **Libyan Telecom Customer Satisfaction Corpus** is a dataset collected from Twitter, focusing on customer opinions regarding the three major telecommunication companies in Libya: **Libyana**, **Almadar Aljadid**, and **Libya Telecom and Technology (LTT)**. This dataset was created to address the lack of sentiment analysis resources for the Libyan dialect in the telecommunication domain. It serves as a benchmark for evaluating machine learning models in predicting customer satisfaction and sentiment polarity (Positive/Negative). The corpus was collected between **January 2020 and December 2020**, capturing customer responses during the COVID-19 pandemic, a period characterized by increased demand for internet services and high competition between providers. 🔗 **Repository:** github.com --- ## 📊 Dataset Statistics ### Raw Collection * **Source:** Twitter API * **Total Raw Tweets:** 16,730 * **Libyana:** 13,782 * **Almadar Aljadid:** 2,551 * **Libya Telecom & Technology:** 397 ### Labeled Dataset Distribution The annotated subset used for experiments contains the following distribution of positive and negative sentiments: | Company | Total Labeled | Positive | Negative | Predominant Sentiment | | :--- | :--- | :--- | :--- | :--- | | **Libyana** | 6,566 | 1,551 (23.62%) | 5,015 (76.37%) | Negative | | **Almadar Aljadid** | 1,784 | 544 (30.5%) | 1,240 (69.5%) | Negative | | **Libya Telecom (LTT)** | 210 | 151 (71.9%) | 59 (28%) | Positive | --- ## 📁 Data Collection & Annotation ### Collection Keywords Tweets were harvested using frequent dialectal hashtags and search terms, including: * `#Libyana` * `#Al-Madar` * `#Libya_Telecom_and_Technology` ### Annotation Process * **Annotators:** Three independent human annotators classified the tweets. * **Method:** Majority vote (the labe …