You're right, and thanks for sharing. It is actually: pip install scikit-ollama

thanks!

Thanks for the comment. While indicators are often used for trend detection, this analysis tests their predictive value for next-day closing prices to evaluate their utility in forecasting. Moving averages are commonly used as direct predictors in ML time series models, but other indicators are less frequently examined this way, which is why we’re testing them here. The shared GitHub code in the video description can be modified to test alternative hypotheses, including different predictive time horizons and relationships to highest highs or lowest lows.

Thanks!

Posted now!

Thanks for the message. The github link is now available in the video description.

@@DeepCharts thanks for the upload, it helps a lot. One more thing: you have clear teaching skills, it’s quite rare among youtubers.

Thanks, the goal of the channel is to be direct and to the point!

Good to know!

Thanks for the message! The code is now up in the video description

I believe that Ploomber has a "secrets" tab upon upload where you can securely place hidden API keys

Thanks!

You could host this app on the Streamlit community cloud for free and access it on your phone. The necessary change you would need to make is to add an input box for your Replicate API token instead of keeping it as a secret variable, so random internet passerby's won't spend your money. Also, check out the channel in a few weeks, I might have an answer for the end of your question.

I won't! Glad you enjoyed it!

delta f"{change:.2f}, change it for: delta=f"{change:.2f}

Thanks, glad you enjoyed it!

Thanks!

So many to potentially learn, but I think each has its strengths for different purposes.

Glad you liked it!

Glad you enjoyed it!

Glad you enjoyed it! Great suggestion!

Glad you find it helpful!

me too

Thanks! And good day to request this since I'll be posting a new one soon.

Pytorch is a great choice. What I've presented here is just an example workflow and methodology. I would test different LLM models for sentiment tasks and different methodologies for turning sentiment into quantitative measures to feed a more comprehensive model. Always make sure to validate your AI measures, as well as your quant finance models, more generally.

Thanks. Glad you liked it!

Thank you!

Glad it was helpful!

It absolutely is!

Glad you like it!

Great to hear!

Coordinates: yes, as long as you have measures of air quality for specific coordinates. Population: no. Population density is often correlated with air quality levels, but kriging doesn't use auxiliary data.

@@DeepCharts Do you know which models use auxiliary variables?

Thanks! Posted now!

Thanks! That was the goal.

Not going to repeat my other comment but any other suggestions for a Business major looking to create such a tool. I dont fit in the quant nor the algo community as I just want information relevant to my trading theory, if that makes sense. Less interested in price prediction. I fear the "it works until it doesnt". Dont know what i dont know but it took months to decide what language to start learning.

Not sure if you watched the video or even read the description. I pretty clearly mention that there are a lot more drivers of stock prices and that this tutorial is just about how to create an LLM workflow that could be used to predict stock prices if more consideration were put into the predictors and the final time series model.

@@DeepCharts Right. That is great. It is just profoundly naive to think that it is useful in any way.

Thank you! More bite-sized geospatial and data science tutorials are on their way.

Thanks!