Arsalan

has anyone tried Bruin Ingestr to get data from #kafka to #bigquery

I want to start an ecommerce company selling buckets. All the no-code businesses of the past few decades are about to kick the bucket, and my goal is to help them with their last purchase.

Every data team needs an AI agent that can act as their full-stack data specialist. We are excited to announce that Bruin is now publicly available for anyone to try. This agent is not just an AI data analyst or AI data engineer - it is an entire data team that lives & breathes inside your organization. Connect your data, ask the agent to analyze data or build dashboards directly from your Slack, Teams, or WhatsApp. The agent can even monitor, manage, and troubleshoot your data pipelines (ingestion, transformation, quality checks, orchestration, governance). Check out our Product Hunt launch: producthunt.com/posts/bruin

In SQL, if you are using MERGE to insert only the delta/incremental data into a table, you should be careful - in some databases like BigQuery, NULL=NULL is FALSE. This means that if a column you are using to merge on has nulls, those rows will be inserted every time the query runs and exponentially grow the table. Make sure you always have quality checks to check for table explosions and null checks.

@samgrows I created a bunch of tutorial videos but needed to convert them to text tutorials for our website, I ended up giving the URLs to Claude and it used an opensource cli tool to fetch the transcription of the videos and convert them into webpages. It was pretty cool.

BREAKING: Claude can now watch videos for you! You can plug in a YouTube, Tiktok or Instagram URL and let Claude break it down for you Live now inside of Algrow

There was the Paris Polymarket manipulator, is there one in Toronto too? @CBC @CTVToronto @CP24

Don't forget to give the repo a star ⭐️ github.com/bruin-data/dac

Our Dashboard-as-Code (DAC) tool is now free and open-source Traditional BI tools are dying because there's no easy way to connect them to AI agents. While those BI companies were seeking no/low code interfaces, the world moved on and code/file based tools are now more powerful because of AI. Bruin DAC is the open standard framework for AI dashboard building and analysis. The It has a built-in semantic layer, it comes with its own skill, and it can connect to all major databases or warehouses: Postgres, MySQL, SQL Server, Snowflake, BigQuery, Databricks, and more. Everything is YAML/JSX based - which also means you can do load-time dynamism (display different charts per value, create tabs or charts dynamically, and have your dashboard live and breathe). We are open-sourcing this tool and it has some rough edges - go ahead and play around with it, give it a try, and share your feedback or contribute to the repo.

Open-sourcing our dashboard-as-code product: dac 🚀 Dac is an open-source product that allows you to build dashboards using YAML or JSX. You can throw your agents at it and get a beautiful, standardized dashboard as a result of it. One of the things I hate about the existing BI tools is that they lock you into a weird UI with no way to bring agents into it. Agents need regular files, but your dashboarding tool doesn't work that way. If you are like me, your first instinct would be to get your agents to build you dashboards from scratch. That's nice to get the first dopamine hit, but it doesn't work well within a team. Every dashboard is different, deploying the dashboards requires a lot of automation work, and there's no way to standardize, govern or review dashboards. Agents generate a ton of code, and it is very hard to review. Enter dac: it is an open standard, and an implementation of the standard as a standalone Go binary. It allows you to build and serve dashboards locally or in any server that can run a single binary. It has a built-in semantic layer, it comes with its own skill, and your agent can build your first dashboard in minutes. Dac supports all the major cloud databases: Postgres, MySQL, SQL Server, Snowflake, BigQuery, Databricks, and more. Another fancy feature of dac is: since it supports JSX, it means you can do load-time dynamism: display different charts per value, create tabs or charts dynamically, and have your dashboard live and breathe. You can customize themes, change the color palette, and make it yours. We are open-sourcing the spec and the implementation today. It has rough edges, but getting started with it should be as simple as "dac init". We are very excited to launch dac and we are looking forward to your feedback on this!

Dashboards as code, finally. Write YAML. Run dac serve. Get a live dashboard in your browser. Bruin dac keeps dashboards next to your SQL, in Git, and querying your warehouse directly. Try it: github.com/bruin-data/dac

Read the whole thing: open.substack.com/pub/arsalandat…

@techbromemes poor claude just wants to listen to some tunes while coding

@tom_doerr posting this on a Friday eh some of us will wake up tomorrow to prod database having been deleted

Autonomous coding agent for relentless self-improvement github.com/uditgoenka/aut…

sql walked into a bar, asked two tables if he join them they said "nice pun nerd", beat sql up and executed him in the back row alley

I always hated using the google, boto3, psycopg2, etc. libraries to connect to a data platform, so many lines of code to do one thing. I've switched over to using bruin's python library and ffs why didn't I start using this earlier

Years ago I set up a Python script to send an email with a csv, just for the team members to then send it via text to another business unit. @getbruin's new Scheduled Reports with seamless integrations into most messaging apps is a game changer⚡ #data #analytics #ai

Your AI data analyst just moved into WhatsApp. Voice messages, file uploads, follow-ups. Ask anything on the go, during or shortly before a meeting. Bruin answers in seconds with tables, charts, and exportable PDFs, right in the chat. <30-second setup:

which one is nicer? #sql #data #duckdb

by the time I started using copilot instead of copy pasting from chatgpt, everyone was already moving onto cursor by the time I started using cursor everyone was on claude code by the time I started using claude code, everyone was using conductor what's next?