Data Donation with Dona: De-identified Messaging Data (WhatsApp and Facebook) and Evaluation Responses

General information

The dataset contains de-identified messaging meta-data from 78 WhatsApp and 7 Facebook data donations. The dataset was collected in an online study using the data donation platform Dona. After donating their messaging data, the study participants viewed visual summaries of their messaging data and evaluated this visual feedback. The responses to the evaluation questions and the sociodemographic data of the participants are also included in the dataset.  

The data was collected from August 2022 to June 2024. 

For more information on Dona, the associated publications and updates, please visit https://mbp-lab.github.io/dona-blog/. 

File description

donation_table.csv - contains general information about the donations including

donation_id: donation identifier

donor_id: the ID of the donor to distinguish the messages sent by them from those sent by contacts

source: the messaging platform from which the data is donated (WhatsApp or Facebook)

external_id: ID used to connect messaging data with the survey data

 messages_table.csv - contains the donated messages including

conversation_id: chat identifier

sender_id: sender identifier

datetime: time of the message, UNIX time for Facebook and device time for WhatsApp

word_count: word count of the messages achieved by splitting the text based on whitespace

donation_id: donation identifier (also listed in donation_table.csv)

messages_filtered_table.csv - same structure as messages_table.csv except that chats with no considerable interactions were removed. This was defined as chats where donor's word count contribution was less than 10% or more than 90%. 

survey.xlsx → contains survey responses of the participants.

survey_table_coding.xlsx → contains the mapping between the column names in survey.xlsx and their meaning, including the original survey questions and response options. Different sheets of the Excel file detail the survey questions and responses in one of the study languages (English, German, Armenian).