October 2020 Summaries

In the second installment of a three-part blog series about building a StackOverflow dashboard using BigQuery and Superset, the focus is on preparing the StackOverflow dataset for visualization and creating basic charts. It explains how to manage the large size of the dataset by creating a filtered view of the questions table in BigQuery, particularly focusing on questions tagged with "superset" or "apache-superset." The post details the process of saving this view and using SQL Lab in Superset to preview it. It introduces the creation of calculated columns to enhance visualization, such as transforming the answer count into a binary classification to identify interesting questions and generating HTML links to questions. The guide also covers the creation of various charts, including Big Number, Big Number with Trendline, Pie, Table, and Box Plot charts, to display data trends, proportions of unanswered questions, and popular questions by view and answer counts. Additionally, it advises on aesthetically enhancing the dashboard with the StackOverflow logo using a Markdown component, and hints at the next part of the series, which will explore more advanced features like joining multiple tables and adding dynamic filters.

Superset has transitioned from relying on auto-generated CRUD endpoints provided by Flask App Builder to developing a new, public, and rigorous REST API designed to support modern React components. This REST API, built with the Marshmallow library for schema definition and Rison for GET requests, is public, versioned, and change-managed, offering better unit test coverage than its predecessor. The move aims to gradually replace the legacy private API with this more refined system, enhancing the quality and consistency of endpoints. As the project progresses, there is potential for a GraphQL implementation to complement the REST API, and future plans may involve making backward-incompatible changes, leading to the development of api/v2. The community is encouraged to engage with Superset’s API advancements, share their usage experiences, and explore the Preset API for programmatic access to various analytics features.

This blog post, the second in a three-part series, guides readers through setting up a basic Slack dashboard using Meltano and Superset. It explains how to load Slack data into a PostgreSQL database using Meltano and then configure Superset to create visualizations from this data. The post provides instructions for installing the PostgreSQL loader, setting up a database, and configuring connection details via a .env file. Once the data is loaded into PostgreSQL, users can register the database with Superset and explore the data in SQL Lab. The post offers a step-by-step process for creating simple visualizations, such as Big Number charts and Tables, which illustrate user metrics and popular time zones, and includes instructions for finalizing the dashboard by adding the Slack logo through a custom Markdown component. The narrative emphasizes the ease of creating charts in Superset without coding and hints at more advanced visualizations and dashboard customizations to be covered in the next installment.