נושא הפרוייקט
מספר פרוייקט
מחלקה
שמות סטודנטים
אימייל
שמות מנחים
המרת משתמשים באפליקצית מסחר בורסאי מקוון
User conversion on an online trading app
תקציר בעיברית
המרת משתמשים באפליקצית מסחר בורסאי מקוון
תקציר באנגלית
The project is made in collaboration with eToro, an online social trading platform in which users trade stocks, crypto, commodities and more. We set our goal to create a statistical model that will help eToro identify the users that are most likely to deposit funds into the app based on their event data (user actions in the app) and therefore improve the conversion rate of registered users to depositors (FTDs). The methodology included 3 main phases: The first revolved around exploring the raw data, making insights, and enriching the original dataset. Then, we implemented a basic model to serve as a benchmark, to which we compared other, more complex machine learning models in the final phase. The end product is a machine learning model that beat the benchmark and performed better than the other models. It has high precision, which is crucial to the business use case as directed by eToro and can be used in two different approaches that are based on different datasets, depending on the use of the company. We can conclude that any model should be made considering the business needs and use cases. Event data is a gold mine which should be utilized to improving the business with predictive models.