@Saanvi_dhillon So, it's not even that models disagree with each other. They disagree with themselves. Because they're non-deterministic. Amazing, but good to know how they behave, to avoid problems elsewhere. If I asked the same question again I'd likely get a different answer.
@Saanvi_dhillon I tried it with Claude Opus 5, ChatGPT, Grok and Perplexity. The answers all differ from what you posted. I talk to people about AI all the time and the question is often "why can't we just have AI do this?" Without guardrails, you get variance just like this.
That moment when you're in a meeting and your boss starts to probe your report in depth, asking for details you don't have? You start to sweat. You feel like you're gonna be fed to a tiger... Now you don't have to worry. You've got Tableau and you saved the world from zombies.
Zombies ate the last analyst on earth like tigers ate Carole's ex...now itโs up to you to learn Tableau and stop them. Join our Tableau Zombie Course to learn everything you need to Tableau like a pro & up your #dataviz skills for under $15 w/code DATAFAM. https://t.co/hnunGzth7P
@SophieWarnes I think dashboards are incredibly useful - if they are done right. And most aren't. Most provide no insight at all and require the end user to analyze everything for themselves. But it's not the dashboard software that is the problem. It's the analyst who hasn't been welltrained.
@pablosdt I think there's lots of interesting stuff going down on Kaggle, but the visualizations, if there are any, could use some Tableau polish. Or build a viz from scratch based on their data and models.Also, you could check out data from Google Data Search or the GDelt project for news
I stumbled across a webinar from Copyblogger this morning about website personalization. If you own a business and are thinking of how to evolve, this might be useful. Learn how to double your list size (and sales) by personalizing your landing pages https://t.co/bLPOyYKzPs