MyGov Review Dataset Construction Based on User Suggestions on MyGov Portal

MyGov (www.mygov.in) is a well known citizen engagement platform (CEP) launched in July 2014 by Government of India (GoI) where millions of people share their innovative ideas about different government services and schemes [1]. The aim of this study is to extract user’s comments, opinion and reviews from MyGov website. As a result, we have constructed two sets of dataset that is labelled and unlabelled user reviews. The labelled one contains 2000 user comments where class label 1 represents useful information for Indian government, 0 represents neutral comments, and -1 category represents comments that contain indignity. The other dataset i.e. unlabelled dataset contains 14,912 user comments in six languages. We have utilized python packages to extract the data that is capable of facilitating the process of fetching information from MyGov, allowing for easy public knowledge of the comments that exist in specific period and serving as a source of inspiration for subsequent development. MyGov has given citizens access to a fresh facet of democracy. By empowering citizens to come up with solutions and join an integrative and participatory governance framework by sharing their perspectives on choices, policies, and programmes made by the government, it promotes the crowdsourcing of ideas from communities. Further, we present classification summary that includes the results of typical machine learning (ML) and deep learning models in single table. In turn, the acquired dataset will be a boon to government to set up policies and benchmark to research community for experimental evaluations.