2018年中国主要城市公共空间社交情感分类比例数据集

新浪微博签到反映了城市中个体的人地关系，个体的微博签到聚合后即可代表城市中居民在不同地点的活动状况，基于此可以用微博签到来量化反演城市公共空间的使用程度，并根据微博签到中的文本进行情感分析计算用户的情感状态。基于2018年中国微博签到数据，根据微博POI进行分类并划分出城市开放公共空间类型，使用BERT深度学习方法进行微博文本情感分类，将定性的微博文本转换为定量化的情感状态概率，从而计算每条签到微博的正面、负面情感概率，最终汇总得出各城市公共空间的正面情感比例。

Sina Weibo check-ins reflect the human-place relationships of individuals in cities. When aggregated, individual Weibo check-ins can represent the activity patterns of residents in different locations within a city. Based on this, Weibo check-ins can be used to quantitatively invert the usage degree of urban public spaces, and sentiment analysis can be performed on the text in Weibo check-ins to calculate users' emotional states. Based on 2018 Chinese Weibo check-in data, classification is carried out according to Weibo POIs to divide into types of urban open public spaces. The BERT deep learning method is used for sentiment classification of Weibo text, converting qualitative Weibo text into quantitative emotional state probabilities, thereby calculating the positive and negative sentiment probabilities for each check-in Weibo post, and finally aggregating to obtain the positive sentiment ratio of each urban public space.