Occupants’ satisfaction with LEED- and non-LEED-certified apartments using social media data
Occupants’ satisfaction with LEED- and non-LEED-certified apartments using social media data
复制标题
使用社交媒体数据评估居住者对 LEED 和非 LEED 认证公寓的满意度
DOI:
10.1016/j.buildenv.2021.108288
复制
发表时间:
2021
影响因子:
7.4
通讯作者:
Liu, Shichao
中科院分区:
文献类型:
--
作者:
Guo, Xingtong;Lee, Kyumin;Wang, Zhe;Liu, Shichao
Leadership in Energy and Environmental Design (LEED) certified buildings aim to offer a sustainable and healthy building environment. Previous studies have shown mixed and inconsistent results on whether occupants in LEED-certified buildings are more satisfied than in non-LEED-certified counterparts. Those studies are usually based on surveys or questionnaires for commercial buildings and were limited by sample size and pre-defined question structures. Since most people would spend more time at home after experiencing the COVID-19 pandemic due to the flexibility to work remotely, assessing the satisfaction with residential buildings benefits future environmental design and certification system development. In this work, we propose a natural language processing-based approach for such assessment. The study collected 16,761 online reviews of 232 LEED-certified and 129 non-LEED-certified apartment buildings from social media, then applied topic modeling and sentiment analysis to evaluate occupants’ satisfaction. Based on topic modeling, we categorized online comments into three topics, 1)location and transportation, 2)running cost, and 3)health andwellbeing.The subsequent sentiment analysis has shown a statistically significant but small or negligible enhancement in the satisfaction occurring in LEED-certified apartments compared to non-LEED-certified ones concerning all three topics. The “significant but small or negligible uptick”has also been found in online star rating and indoor environmental satisfaction. The only exception with a large effect size is lighting which is found to be significantly more satisfying in LEED-certified apartments. Nevertheless, the statistical significance in online star rating disappears when normalized by rent price and property house value.