LangChain In Action: Real-World Use Case With Step-by-Step Tutorial

In this video, we are going to have a look at a real-world practical application of Langchain. We're going to build a small Customer Experience Analytics Python library and have GPT-4 analyze the customer reviews using Pinecone as a vector store and Hugging Face embeddings.

The data used in the video is available here:

The code for the video is available here:

▬▬▬▬▬▬ V I D E O C H A P T E R S & T I M E S T A M P S ▬▬▬▬▬▬

0:00 Introduction and overview

1:35 The Amazon review data

3:24 Creating the embedding vectors and finding signal in review data

6:10 Combining GPT-4 with Pinecone vector storage

9:00 Doing filtered vector similarity search with Pinecone

[jarrodhroberson]

Very nice this is the first practical example of using these tools to do with a good at instead of what they’re hyped and not good at thank you

[crisgath]

Thanks for the content, your channel is on point with production themed ideas....

[rishabhupadhye]

This was wonderful! Looking forward to more such use-cases

[IamMarcusTurner]

that was great looking at more solutions around structured data so thank you for this tutorial

[umeshtiwari]

great use case. Hats off. kindly make more more uses. i can image many people can watch your video and land some good job oppotunities. i really liked it. Thanks again

[cubibiris]

I really love your videos, just amazing work. Thanks.

[VietTranIT]

Thanks for the great tutorial/introduction. Keep on the good work.

[shaheerzaman]

you can store text also as metadata. that way you get to avoid a local copy of the data.

[angelo]

This is wonderful. How long would creating this app take? you made it look easy!

[jaivalani]

Thank you ! One q row: dict(rating=row['overall']), axis=1) what values doe sthis leads to. In one your sections this field tend to had different values like genre and that.How does it builds it Pinecone

[andreshofmann]

anyone else having issues with sentence_transformers? even when I installed it in many ways I keep getting an import error

[sandorkonya]

Could we use image embedings in vectore stores? Like... i have a text book (pdf) full with illustrations and figures, and i would like to search for images based on an query image?

[wryltxw]

Can OpenAI understand HuggingFace Embeddings?

[atsileitumeleng]

Can you please cover the cost per user interaction if this is released out as a consumable application to users?

[brianconnolly]

What value does pinecone add here compared to the raw embeddings when running the "stuff" chain? I'm not clear what it's actually doing. Thanks for the tutorial.

[veliea]

Why did he train data with sklearn? did he create a vector store in memory with sklearn?

[ipxify]

do you have a coure to buy? need your knowledge !

[mweishuhn]

If you are using a product that has over 4300 reviews, truncating them at 400 characters a piece, how are you then passing all of those reviews into gpt4 for analysis without hitting the token limit?

[kelvincht]

This video is mostly about setting up data rather than using Langchain features such as Agents and chaining. If you decided to have another video, please make this video part 1 of n

[BOGABOOfull]

Awesome video :) Is this possible without using paid models, such as gpt 4? Are there any free "chat type" models available in langchain?