Urban forests can attract visitors by the function of well-being improvement, which can be evaluated by analyzing the big-data from the social networking services (SNS). In this study, 935 facial images of visitors to nine urban forest parks were screened and downloaded from check-in records in the SNS platform of Sina Micro-Blog at cities of Changchun, Harbin, and Shenyang in Northeast China. Images were recognized for facial expressions by FaceReader TM to read out eight emotional expressions: neutral, happy, sad, angry, surprised, scared, disgusted, and contempt. The number of images by women was larger than that by men. Compared to images from Changchun, those from Shenyang harbored higher neutral degree, which showed a positive relationship with the distance of forest park from downtown. In Changchun, the angry, surprised, and disgusted degrees decreased with the increase of distance of forest park from downtown, while the happy and disgusted degrees showed the same trend in Shenyang. In forest parks at city center and remote-rural areas, the neutral degree was positively correlated with the angry, surprised and contempt degrees but negatively correlated with the happy and disgusted degrees. In the sub-urban area the correlation of neutral with both surprised and disgusted degrees disappeared. Our study can be referred to by urban planning to evaluate the perceived well-being in urban forests through analyzing facial expressions of images from SNS.PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.3424v1 | CC BY 4.0 Open Access |

Urban forests can attract visitors by the function of well-being improvement, which can be evaluated by analyzing the big-data from the social networking services (SNS). In this study, 935 facial images of visitors to nine urban forest parks were screened and downloaded from check-in records in the SNS platform of Sina Micro-Blog at cities of Changchun, Harbin, and Shenyang in Northeast China. Images were recognized for facial expressions by FaceReader TM to read out eight emotional expressions: neutral, happy, sad, angry, surprised, scared, disgusted, and contempt. The number of images by women was larger than that by men. Compared to images from Changchun, those from Shenyang harbored higher neutral degree, which showed a positive relationship with the distance of forest park from downtown. In Changchun, the angry, surprised, and disgusted degrees decreased with the increase of distance of forest park from downtown, while the happy and disgusted degrees showed the same trend in Shenyang. In forest parks at city center and remote-rural areas, the neutral degree was positively correlated with the angry, surprised and contempt degrees but negatively correlated with the happy and disgusted degrees. In the sub-urban area the correlation of neutral with both surprised and disgusted degrees disappeared. Our study can be referred to by urban planning to evaluate the perceived well-being in urban forests through analyzing facial expressions of images from SNS.PeerJ Preprints | https://doi.org/10.7287/peerj.preprints.3424v1 | CC BY 4.0 Open Access |