The COVID-19 pandemic, which was officially recognized by the World Health Organization in March 2020, has caused remarkable changes in global public attitudes, economic environments, and cultural dynamics. It is important to understand the diverse reactions and feelings of people towards the pandemic in order to inform effective public health initiatives, policy interventions, and social reactions. Social media platforms have become valuable real-time data libraries, providing insights into people's emotional reactions and general awareness during times of crisis. The availability of datasets with social media content relevant to COVID-19 becomes significant in this scenario. The dataset available at https://www.kaggle.com/datasets/valbauman/canada-covid19-tweets is a useful tool for researching viewpoints on the epidemic that have been made in Canada. Although the dataset is limited to tweets about COVID-19 in Canada, it is a helpful starting point for examining regional attitudes and patterns. SenWave's analytical approach, when combined with this dataset, allows for a focused look into sentiment dynamics among Canadian citizens. Through the application of machine learning methodologies and sentiment analysis algorithms, researchers may explore the complexity of public opinion, classifying expressions into several sentiment categories including optimism, thankfulness, anxiety etc In short,the research paper's approach, when combined with the dataset given, allows for a deeper investigation of Canadian attitudes toward COVID-19 in the context of global sentiment dynamics. This multidisciplinary approach has the potential to guide proven strategies and treatments that target public concerns, build adaptability, and promote unity among individuals in overcoming the obstacles this ongoing pandemic presents.
We trained the model Using BERT on a different dataset to re evaluate the dataset. After that we decided to take 'Canada Covid19 Tweets' dataset from Kaggle published by 'Val Bauman'.
(https://www.kaggle.com/datasets/valbauman/canada-covid19-tweets)We just cleaned the data by removing stopwords, punctuations, performed lemmatization etc. we also created abbrevations and word replace text document and used this file in the data cleaning file. used function to take these two files and performed adding abbrevatioins and word replace. After preprocessing created and saved the preprocessed csv file
we used preprocessed dataset for training and building model. found sentiment score, emotions, emotion analyzed based on tweets. used word cloud and displayed emotions.
In order to examine localized sentiment dynamics, this study combined the SenWave framework with a collection of Canadian tweets pertaining to COVID-19. Despite many limitations, including regional restrictions and linguistic prejudices, the study yielded insightful information about sentiment patterns in Canada. Further studies may examine other data sources and tackle ethical issues. All things considered, this effort enhances knowledge of the worldwide sentiment patterns during the COVID-19 pandemic and supports evidence-based approaches to resolving regional differences in sentiment expression in Canada.Incorporating geolocation data to study sentiment patterns at the city level could be a future development, building upon the existing dataset structure and the framework described in the research paper. Researchers can obtain insights into sentiment changes across various Canadian cities and regions by precisely geotagging tweets. Furthermore, the incorporation of user demographic data, including age, gender, and occupation, may augment the comprehension of sentiment dynamics and facilitate focused outreach initiatives. By combining sentiment analysis with geolocation data and user demographics, further improvements can provide a more detailed and localized view on Canadian sentiment trends during the COVID-19 epidemic, enabling targeted and effective responses
SENWAVE: MONITORING THE GLOBAL SENTIMENTS UNDER THE COVID-19 PANDEMIC
https://arxiv.org/pdf/2006.10842.pdf