crowdsourcing-data

Dataset Overview: User Task Participation and Social Fraud DetectionThis dataset contains records of user task participation, social network relationships, and related fraud detection information. It is designed for analyzing user behavior patterns, social network influences, and task completion outcomes. The data is stored in CSV format across 7 files.1. File Descriptions and Rolesfinal_data_all_with_result.csv (Master Data): This is the final integrated dataset. It contains all user task participation information consolidated into a single file, along with outcome labels such as fraud indicators and completion status.friend.csv (Social Relationship): This file maps the friendship connections between users. It typically includes user IDs and their corresponding friend IDs, which are essential for constructing social network graphs.msg.csv (Communication Log): This contains the message exchange records between users. It includes details such as sender IDs, receiver IDs, and timestamps to track interaction frequency and timing.nx_participation_temp.csv (Intermediate Data): This is a temporary file generated using the NetworkX library. It is likely used for graph analysis or as a cache for specific feature engineering steps.nx_task_fraud.csv (Task Fraud): This file focuses on task-level fraud detection results. It contains task IDs paired with specific fraud metrics or risk scores.participation.csv (Participation Record): These are the core records of user activity. Each entry tracks a user’s involvement in a specific task, including the user ID, task ID, and the time of participation.task.csv (Task Information): This file defines the attributes of the tasks themselves, such as the task name, release time, and the reward amount offered for completion.2. Data Relationships and IntegrationTask Linking: The participation.csv file can be linked with task.csv via the Task ID to retrieve detailed attributes for every user action.Social Network Construction: friend.csv and msg.csv are used to build the user social graph. When combined with participation records, this allows for the analysis of social influence and group behavior.Consolidated Variables: final_data_all_with_result.csv serves as the primary integrated source, likely derived from merging the other files to include all final target variables (labels).Graph Computation: Files prefixed with nx_ are intermediate processing files, specifically optimized for graph algorithm computations or temporary storage of network-based features.3. Usage Suggestions and ToolsInitial Analysis: It is recommended to start your analysis with final_data_all_with_result.csv, as it provides the most comprehensive view of the data immediately.Graph Modeling: To build a user social network graph or detect "collusion" clusters, utilize friend.csv or msg.csv.Task-Level Insights: For analyzing which tasks are most prone to fraud, link task.csv with nx_task_fraud.csv.Technical Stack: We suggest using the Python pandas library for data manipulation and merging, and the NetworkX library for conducting complex social network analysis.