Power BI

“you can skip to 2:00 to see the results, But this was made using single prompt. Things like this used to be made in hours, now in minutes Edit- i am using this method from last 5 months, working on PBIP template (the initital github link provided in prompt), yes it is ready for production business data. First, you use Python to do EDA of the DB and feed the business/data context; next, focus on building a Power BI dashboard. I am making a hour long tutorial on this using production sales data re”

“Sadly, another fairly weak month. I know the MSFT team cares deeply, and I do very much root for them, but there is a persistent refusal to address even the most basic Quality of Life improvements for which the PBI community has been begging for literally years: Designing reports is exceedingly and needlessly complex, core visuals like Matrices still have extremely limited formatting, shoehorned AI features that don't justify the price, and a philosophy towards Excel integration that can only be”

“I'll keep this brief. With Claude, we can now build dashboards and data apps locally, quickly and easily, with far more freedom, and connected directly to Snowflake. In my 10 years as a data engineer, I've never used anything as interactive or advanced as these apps. The future of data analysis: ERP + Data Warehouse + centralized local apps.”

“A few thoughts Yes, MCP servers are a game changer - this has potential to change how we interact with data. Claude with power bi MCP server is a real eye opener The data need to stay regional in most corporate environments and running a local model as capable as Claudes models is not feasible. It's also stupid expensive and most ai runs at a financial loss - so there are currently practical limits around corporate usage Power BI semantic models are the real valuable part of fabric, and da”

“I spoke to a lot of Qlik and Tableau guys recently as our company is moving to PowerBI across the board. The Qlik guys love their product, but their main reason for getting with PowerBI is cost (it’s free with an E5 license from MS, which everyone in our company has), and because they can’t backfill Qlik vacancies. They just can’t find the people, so they conclude the product is effectively dead if they can’t find people to develop in it. We will reduce spend by somewhere close to 7 figures when”

“Yeah until things go wrong. This week my company had to spend $120k on a new P2 premium capacity, solely because 95% analyst can't build performant data models that align with best practices.”

“Sadly, another fairly weak month. I know the MSFT team cares deeply, and I do very much root for them, but there is a persistent refusal to address even the most basic Quality of Life improvements for which the PBI community has been begging for literally years: Designing reports is exceedingly and needlessly complex, core visuals like Matrices still have extremely limited formatting, shoehorned AI features that don't justify the price, and a philosophy towards Excel integration that can only be”

“Microsoft Power BI is the best BI tool by far | it is the best tool out there, the desktop version is free to use so you can use it in your own projects, it is also appropriate for very large organizations because it has elastic scalability plans, it also well suited for complex scenarios where you need to have everything in one place with the use of fabric. it is not suited if you have very basic reporting and not really large datasets. | Pros: StoryTelling | Pros: Report Development | Pros: H”

“Power BI would stay. Your company is soon going to realize the maintenance headache of maintaining all dashboards, and the unreliability of AI assisted development. I have been doing AI assisted development for over a year now including React based applications, dashboards, data engineering and Power BI. AI is incredibly unreliable, even after writing so many test cases. Power BI on the other hand guarantees an output when a certain input is provided.”

“ok, I see a few issues. you may know this but I'm just going off your post Separate your database into a production db and an analytics db Production db data corruption - every power bi user becomes a risk factor if they don't write their custom sql queries correctly. Including missing the 'with (nolock)' hint. This will lock up the production database at best, and corrupt it at worst. Production db processing slowness - the production database is probably a transactional/relational databa”

“Unreliable and inconsistent: Unfortunately, this app can't be relied on to produce what you expect. A dashboard which works in web will have tiles showing an error on iPhone and, on iPad, briefly loads dashboard correctly then becomes a completely blank screen.”

“Awesome app: Great app! Had a problem after installing the latest version and couldn't render any dashboards. Logout/login/kill app, restarted phone, no luck. Had to remove and reinstall.”

“I'll keep this brief. With Claude, we can now build dashboards and data apps locally, quickly and easily, with far more freedom, and connected directly to Snowflake. In my 10 years as a data engineer, I've never used anything as interactive or advanced as these apps. The future of data analysis: ERP + Data Warehouse + centralized local apps.”

“I have tried it and it is extremely ass. Unusable almost. There are many other dashboarding tools that you can write with Ai agents much easier. We've been using a lot of streamlit lately and dbt charts also seems promising. We want to phase pbi out because it's so inflexible and it's just so much faster making up a dashboard with code and an Ai agent.”

“I blame Microsoft for this one in all honesty. I get the need to advertise something as "easy" as it attracts a wider audience but... It isn't easy. I don't think I have ever built anything for a client in the 7 years I've been working with it that has been pure drag and drop all the way from data ingestion to report development. The semantic model is hard to develop properly and DAX is just insane to explain to a non-technical person. If the model is already built properly then it is signific”

“Feel the same. I've been using Power Bi intensively(everyday in a professional capacity) for over a year and... there is so much I still don't know how to do. Someone thinking that anyone can use Power Bi is a lunatic. Copilot within Power Bi is supposed to be able to build a lot for people providing the data model(semantic model??) Is AI friendly. Someone, however, has to create this model and it won't be AI.”