81 5 CHAPTER V CONCLUSION AND RECOMMENDATION 5.1 CONCLUSION The author tested 50% of sentiments from 2,100 data points after pre-processing. Initially, the author collected more than 5,000 tweets. Various test data and data mining processes were conducted, resulting in an analysis based on multiple data findings. Based on the first research question, the positive, neutral, and negative sentiment trends showed similar patterns across two periods: during the pandemic and post-pandemic. No significant difference was observed in sentiment movement between these periods. This indicates that employees consistently and predominantly perceive working from home negatively, viewing it as full of challenges, drawbacks, and obstacles. However, this is followed by some employees who remain neutral, while only a few prefer this work method arrangement. Based on the second research question, a wide distribution of sentiments emerged toward various digital communication tools analyzed in this study, such as Zoom, Slack, Google Meet, and Skype. These were identified as the top tools appearing in the collected sentiments. The sentiment distribution fluctuates—positive, negative, and neutral—depending on individual preferences when using these tools. However, they generally perceive these platforms as standard tools for digital interaction. Rather than focusing on specific features or specifications, Indonesian workers on X tend to express their perspectives on the diverse behaviors of individuals in digital communication and interaction. Thus, while the tools themselves are considered acceptable, the way they are used for interaction remains the dominant concern. Based on the third research question, the author analyzed employee sentiment and found that more than 50% of each period's sentiment was negative toward digital communication during remote work. This indicates that the author's future strategies should begin with identifying negative sentiments. The author discovered a list of top keywords associated with this dominant negative sentiment, including 'Hard,' 'Miscommunication,' 'Tired,' 'Complicated,' 'Children, 'Leave,' and 'Slow Response.' 82 5.2 RECOMMENDATION The research is expected to answer questions about digital communication during and after pandemic situations. To give recommendations, the author analyzes based on the potential challenges employees have faced through the analysis of RQ3 for the business. Additionally, it aims to deepen future researchers' understanding of X data mining and Sentiment analysis from X. 5.2.1 Recommendation for Business Here are some recommendations for Business 1. This research shows that many individuals perceive working from home as challenging, with dominant-negative sentiments.