Dataset on individual differences in self-reported personality and inferred emotional expression in profile pictures of Italian Facebook users

We retrieved the current profile picture of 2234 Italian Facebook users who also answered self-report questionnaires on demographic variables and personality. Data were collected between March and June 2018 using a Facebook wep application. Profile pictures consisting of 200x200 resolution jpegs were obtained by sending a request via the Facebook Graph API and analyzed using online commercial services allowing for the scoring of facial expressions in image data, namely Microsoft Azure Face APIand MEGVII Face++ Detect API. Both services provide emotional expression scores if at least one (n = 1) face is successfully detected in the picture. Using the Microsoft Azure Face API we obtained scores for anger, contempt, disgust, fear, joy, sadness, surprise, and neutrality. Using the MEGVII Face++ API, pictures were scored for the presence of anger, disgust, fear, joy, sadness, and surprise, and neutrality. Higher scores on each emotion refer to a stronger expression of the respective emotion.
The dataset presented here consists of data of N =728 Facebook users with a profile picture in which both APIs detected only one (N=1) face.
Regarding self-report data, the dataset includes the following demographic information about the participants: gender and age. The dataset also includes participants’ personality scores based on a short validated assessment of Big Five traits (Ten Item Personality Inventory), and Impulsivity/Sensation Seeking (IMPSS8). A document included the questions administered in the online survey is attached to the dataset.
This dataset can be useful to generate insights on the association between demographic variables, including age and gender, and personality (Big Five traits and Impulsivity/Sensation Seeking), and emotional expression as derived from social media pictures. It can be useful for researchers and data scientists who do research in social sciences, in particular psychoinformatics, to train models in order to infer personality of users of social media platforms from profile pictures.

The annexed files include the following:
DIB_DATASET_25_10_2021.csv (the actual data)
DIB_DATASET_Codebook.xlsx (the codebook for the data)
Supplementary material - Online survey.docx (doc file including questions administered to participants)

本研究获取了2234名意大利脸书（Facebook）用户的当前头像，这些用户同时填写了人口统计学变量与人格相关的自陈问卷。数据采集于2018年3月至6月期间，通过脸书网页应用完成。所有头像均为200×200分辨率的JPEG图片，通过脸书图形API（Facebook Graph API）发起请求获取，并借助两款支持图像数据面部表情评分的在线商业服务进行分析，即微软Azure面部识别API（Microsoft Azure Face API）与旷视Face++检测API（MEGVII Face++ Detect API）。若头像中成功检测到至少1张人脸（n=1），两款服务均会输出情绪表达评分。借助微软Azure面部识别API，我们获取了愤怒、轻蔑、厌恶、恐惧、快乐、悲伤、惊讶与中性共8类情绪的评分；而通过旷视Face++检测API，我们对头像的愤怒、厌恶、恐惧、快乐、悲伤、惊讶及中性情绪进行了评分。各类情绪的评分越高，对应面部表情的强度越强。

本数据集共包含728名脸书用户的数据，这些用户的头像均被两款API同时检测到且仅含1张人脸（N=1）。

在自陈数据方面，本数据集包含参与者的两项人口统计学信息：性别与年龄。此外，数据集还包含基于经过验证的大五人格（Big Five traits）简版测评——十项人格量表（Ten Item Personality Inventory），以及冲动性/感觉寻求（Impulsivity/Sensation Seeking）量表（IMPSS8）得到的参与者人格评分。本数据集附带一份在线调研使用的问卷文档。

本数据集可用于探索人口统计学变量（含年龄与性别）、人格特征（大五人格与冲动性/感觉寻求）与社交媒体头像衍生的面部情绪表达之间的关联。同时，该数据集可用于帮助社会科学（尤其是心理信息学（psychoinformatics））领域的研究者与数据科学家训练模型，以通过社交媒体用户的头像推断其人格特征。

随数据集附带的文件如下：
DIB_DATASET_25_10_2021.csv（原始数据集文件）
DIB_DATASET_Codebook.xlsx（数据集编码手册）
Supplementary material - Online survey.docx（包含调研问卷的补充材料文档）