In a streaming workload scenario, you are required to process real-time data from social media feeds. The data includes text, user information, and timestamp. What partition strategy would you implement to handle the high volume and velocity of this data effectively?