November 2024 Summaries

The text discusses the importance of curating data for Explore, a no-code experience that allows business partners to explore data directly using visual data exploration without needing to learn code or wait on the data team. To ensure a smooth and relevant experience for these users, it is recommended to set up a new data connection with only the necessary databases, schemas, and tables. This can be achieved through schema filtering in the Data browser, Magic AI include/exclude toggles and endorsements, and configuring user permissions in the actual warehouse. Additionally, adding descriptive metadata to the data helps improve accuracy when using tools like Magic for data exploration.

Hex is integrating its Magic tools with Cortex, a platform for deploying and managing generative AI models. This integration will allow users to generate Python, edit SQL queries, or build no-code data visualizations using natural language prompts. Claude 3.5 Sonnet, an AI model known for its generation and reasoning skills, is also being added to Cortex. Hex users can already use Cortex in their analytical projects by calling functions in SQL and adding the Cortex API to their Python cells. The integration will provide enhanced control, governance, and observability while keeping data within the user's Snowflake account.

Metrics act as a compass for businesses navigating through uncharted waters. As companies evolve, the number of reports and KPIs increase, making it time-consuming to manually check them all. To help cut through the noise, native alerting capabilities have been introduced, which bring important data updates directly to users via push notifications in Slack or email. These alerts can be customized based on specific conditions and thresholds, ensuring that crucial information is delivered promptly. Examples of alerts include KPI tracking, MRR tracking, traffic tracking, CVR tracking, reporting period completion, user behavior monitoring, large discounts tracking, user account creation tracking, changes in viewer seats tracking, test failures tracking, goal completion tracking, and unusual API activity tracking. Alerts help make everyone a data person by keeping business partners up to date on data team work in Hex and allowing them to self-serve answers.

Hex has introduced no-code joins in its recent Explore release, allowing low-code and no-code users to perform joins without needing SQL knowledge or assistance from the data team. The intuitive guardrails ensure accurate results even during fan-outs (duplication of rows). No-code joins combine a point-and-click interface with AI suggestions for join keys and column profiling, making it easier for everyone to handle complex SQL queries. This feature helps keep analysis flowing, enabling business users to answer more intricate data questions confidently while allowing the data team to focus on insights rather than troubleshooting joins.

In an interview with Ramp's Head of Analytics Engineering and Data Science, Ian Macomber, he discussed his philosophy for building a high-performing data team. He emphasized the importance of leveraging data to create unique products and using A/B tests only as a last resort. Instead of measuring everything, he believes in focusing on product thinking and strategic decision making. Regarding AI's impact on data jobs, Ian stated that it will not replace them but may change the nature of some roles. He looks for candidates with curiosity, passion, and adaptability to learn new tools quickly.