Universal –

• Uid: a unique identifier for the user
• Predicted_Gender: refers to the estimated gender of the user
• Predicted_Age: refers to the estimated age bucket of the user
• Login_mode: refers to the authentication details provided by the user such as Facebook, Email, Phone etc
• Streams: refers to the total number of times a song was played by the user. This helps one estimate general stickiness of the user to the application and thus, the probability of the user returning to the application
• Price: refers to the estimated price of the device owned by the user
• City: refers to the location of the user

Language dataset -

• Language: refers to the language of music streamed by the user

Ads dataset –

• Gen_conversion: refers to the % of all ads the user interacted with in the past. For example, if the user received 100 ads and interacted with 5, the total gen_conversion would equal 0.05. This helps one estimate general propensity of the user to engage with advertisements
• Ecomm_conversion: refers to the % of all ecommerce ads the user interacted with in the past . Similar to gen_conversion, if the user received 100 ads from an ecommerce brand and interacted with 5, the total ecomm_conversion would equal 0.05
• Last_clicked: refers to the date of the last click of the user
• Last_active: refers to the date the user was last active on the platform. That is, the last date on which any streaming activity was recorded for the user