04-27, 10:15–11:00 (America/Los_Angeles), Rainier
This talk presents a case-study of replacing a proprietary marketing analytics platform with a dashboard and web app created using the Python data ecosystem. The app will provide the analytics features found in popular paid alternatives in an accessible web interface, and demonstrates how data science teams can be empowered to create and deploy applications which have distinct advantages over commercial alternatives.
Companies often adopt paid SaaS platforms for analytics, which are inflexible, expensive, and encourage the sharing of users' data with third-parties.
In addition to providing world-class libraries for data analysis and visualization, Python provides powerful ways to create dashboards and internal tools. However, it is often seen as a poor choice due to the perceived complexity of turning a data analysis into a full-stack application for non-technical audiences.
This talk will challenge that assumption, and explore the case-study of replacing a commercial marketing analytics platform with a dashboard and web app built using Python’s data ecosystem.
The audience will be provided with a sample analysis which pulls data from an ecommerce platform’s API, and performs methods such as subscription churn prediction, cohort analysis, and LTV prediction using a combination of classical statistical approaches, machine learning, and specialised open-source libraries.
From this analysis, we will demonstrate how to create reusable visualizations, dashboards and reports, and finally a deployed interactive web application which rivals commercial alternatives.
The source code will be available as a Jupyter Notebook and GitHub repo for the audience to use.
No previous knowledge expected
Leo is the co-founder and CEO of Datapane, an open-source framework for creating data apps using Python and Jupyter.