August 2022 Summaries

In early August, Preset announced its Embedded SDK, enabling users to integrate Preset dashboards into external web applications to enhance data application experiences. This blog post serves as a companion to a live demo conducted on August 16th, guiding users through building their first embedded dashboard using the SDK. Users must have access to the embedded feature, familiarity with web application development, and a dashboard to embed. The procedure involves creating a web app with both frontend and backend components, using Python's Flask framework for the backend to request and manage authentication tokens from the Preset API. The frontend, using HTML, CSS, and JavaScript, fetches guest tokens from the backend and utilizes the Preset Embedded SDK to embed the dashboard as an iFrame. The Preset API builds on the Superset API, requiring an API key for interaction, and although currently available by request, offers powerful tools for creating interactive analytics experiences in customer-facing applications.

The modern data stack is a cloud-native framework designed to handle the increasing volume and complexity of data within organizations, facilitating the transition from legacy systems to more agile, scalable solutions like cloud data warehouses. This stack has evolved due to the rise of technologies such as Amazon Redshift, Google BigQuery, and Snowflake, which offer benefits like faster processing speeds, easier connectivity, and greater accessibility compared to traditional on-premise systems. The shift from ETL (extract, transform, load) to ELT (extract, load, transform) has been enabled by the cloud, allowing organizations to store raw data and transform it as needed, reducing costs and strategic constraints during the data ingestion phase. The stack's components—ranging from data integration and transformation tools like Airbyte and dbt to orchestration tools like Airflow and reverse ETL solutions like Hightouch—are designed to democratize data access and support data-driven decision-making across various business functions. Moreover, data observability tools such as Monte Carlo and Datadog monitor data quality, ensuring reliable insights and minimizing the risk of data-related issues. The modern data stack ultimately enhances business intelligence capabilities by improving data accessibility and comprehension, enabling faster, more informed decision-making.

Preset has announced the release of its Embedded SDK, which allows users to integrate interactive dashboards within external applications to enhance analytics experiences for end-users. This SDK enables quick deployment of dashboards with minimal coding, utilizing guest tokens for user access without requiring login. The solution includes over 40 visualization types, allowing for dynamic and customizable dashboards, and integrates into the software development lifecycle as code. Preset, a major contributor to Apache Superset, offers managed services that enhance the Superset experience, providing regular updates, SOC2 compliance, and robust privacy measures. Users can manage resources such as databases and dashboards with powerful APIs, enabling scalability and performance optimization. Preset Cloud facilitates easy migration for existing Superset users, offering a seamless path to embedding dashboards and controlling access with role-based permissions.

Apache Superset and Apache Airflow, both originating from Airbnb, exemplify the company's significant contributions to modern open-source data tools, some of which have led to the creation of independent companies. In a discussion on the Decoding Data Leadership podcast, Max conversed with former Airbnb data experts Chetan Sharma, Lindsay Pettingil, and James Mayfield, along with Swaroop Jagadish from Acryl Data, about various challenges related to building large data organizations and fostering a data-centric culture. They explored topics such as data democratization strategies, managing and rolling out open-source projects, cultivating internal political capital for data, and the future direction of data management.

Max appeared on the MAD Data Podcast, where he discussed a range of topics related to data quality, machine learning, and artificial intelligence. During the two-part episode, Max delved into IBM's acquisition of Databand and the significance of open-source data tooling, alongside exploring the intersection of Data Ops and DevOps. He highlighted the diminishing art of data modeling and analytics and examined the trade-offs between normalizing and denormalizing data. Max also discussed the dataset metaphor as a means to balance human-centered high-level abstractions and computer-optimized lower-level structures, touching on the limitations of semantic layers and the role of Tableau extracts as datasets. Additionally, he explored how Superset supports various approaches, including query-centric, dataset-centric, and semantic-centric methodologies, questioning whether the transform layer adequately captures semantic layer features.

Preset has introduced a range of new features and enhancements designed to improve the user experience across its platform, particularly in charting, dashboarding, and data integration. Notably, Preset now offers Managed Private Cloud (MPC) for deploying workspaces within a private AWS cloud and has integrated with dbt core to streamline database and dataset management. The platform also supports Databricks Partner Connect, enabling easier connectivity between Databricks workspaces and Preset Cloud. New charting capabilities include a quick visualization switcher, one-click chart editing, and enhanced dashboard filters, while visualization enhancements support features like Horizontal Bar Charts and Mixed Time-series charts. Additionally, the Preset API and CLI now facilitate dashboards-as-code workflows, and SQL Lab has been upgraded with a find and replace tool. The Preset Manager allows for customization of logos and color palettes, role assignments for workspace members, and access to query logs, with various UI improvements and functionalities added to support a more seamless and efficient user experience.

Preset offers a user-friendly interface for managing databases, datasets, charts, and dashboards, catering to both technical and non-technical users. The article illustrates how Preset's integration with dbt allows for managing Apache Superset assets as code, encouraging version control, code review, and continuous deployment. It presents a case study involving a bike-sharing company, which utilizes dbt to transform raw trip data into actionable insights through dashboards, helping the company optimize services and increase subscriptions. The process entails connecting databases to Preset, conducting a discovery phase to decide on dashboard content, and using dbt to create transformed datasets that are synchronized with Preset for visualization. The integration of Preset with dbt enhances the ability to manage data and dashboards systematically, as assets can be exported as YAML files, enabling source control and collaborative management. This approach ensures insights are built on a trustworthy data infrastructure, supporting better decision-making and organizational efficiency.