Several solutions have been created for automating specific data science scenarios and implementations of personalized content in conversational interfaces. However, the overall understanding of these conversational interfaces that provide personalized suggestions for data scientists is still poorly explored. We identify the need to automate data science procedures up to different levels of automation. Our research focuses on helping data scientists during the automation of these procedures by using conversational interfaces. We propose a framework for creating a chat-bot system to facilitate the automation of data science common scenarios. In addition, we instantiate the framework in two different data science scenarios. The first scenario focuses on outlier detection, and the second scenario on data cleaning. We conducted a study with 28 participants to demonstrate that data scientists can use the proposed framework. All participants completed the activities correctly, and 75 to 80 percent found the framework relatively easy to extend and use. Our analysis suggests that the use of conversational interfaces can facilitate the automation of data science tasks.