Build a Python Web App to chat with your datasets using GenAI

🤖 In this exciting episode, Andrianasolo Fanilo takes you on a journey into the world of artificial intelligence and data exploration. #TechAtWorldline #PythonWebApp #GenAI #DataChatbot #DataExploration #AIInData
📊 Get ready to learn how to create a Python web application that will allow you to have interactive conversations with your datasets using using Streamlit & PandasAI.
🧪 Checkout the source code here :
github.com/worldline/streamli...
👩‍💻 Whether you're a data scientist, developer, or just curious about the intersection of AI and data, this tutorial is for you. Join us on this exciting journey and unlock the full potential of your data.
👍 If you find this video helpful, don't forget to give it a thumbs up and share it with your fellow tech enthusiasts. Have any questions or feedback? Leave a comment below; we'd love to hear from you.
🔗 Stay tuned for more exciting tech content by subscribing to Tech at Worldline, and don't forget to hit the notification bell so you never miss an update!
👇 Get more content from Fanilo on his channel
@Fanilo Andrianasolo
Follow us:
🌍 blog.worldline.tech/
/ worldlinetech

[@natawebmaster]

Thanks for video. Tomorrow will try.

[@mikew2883]

Awesome! 👍

[@andfanilo]

Thanks for the feedback :)

[@nikhil182]

Can't believe you have only 525 subscribers! The content is really good.

[@andfanilo]

Thank you. I have a personal channel with similar content if you want to check it out :)

[@WhySoBroke]

Very cool!! Any updates or new discoveries of chat with data methodologies??

[@yrahaman7199]

Love the video ❤.
Could you please do it using langchain once

[@akshitsinghal8590]

I really like this project will work on this coming weekend.

[@andfanilo]

Awesome, good luck 🙂

[@GeriFitrah]

I have questions, can it run follow up question ? Or it only work once and done ? Ex i want you show data 2023, follow up question show only Adams data ?

[@stephenpire3803]

Amazing, I'm a huge fan of your channel, and learned a lot thanks to you! However I got an issue, for some reason, the temp file of the graph is correctly generated (and stored in ./exports/charts/) however, it doesn't show up in Streamlit. Instead, it's displayed as a doc icon with a 0. Any idea what I might have done wrong?

[@andfanilo]

Hey there! That's strange, I just cloned the project and I have the same problem.
I ended up reloading the image with Pillow and passing this to st.image instead of passing the file path to st.image
so:
```
from PIL import Image
class StreamlitResponse(ResponseParser):
def format_plot(self, result):
img = Image.open(result["value"])
st.image(img)
return
```
Pushing the changes to the repo :) (and also pinning pandasi as apparently callbacks have disappeared since the release of the video), thanks for sharing!

[@nunomaltez804]

thnk you, summary of what i needed.

[@andfanilo]

Glad to hear it was useful 🙂

[@vinaya68vinno1]

Images not loading in pickle format can you explain this video brief

[@KALYAN1898]

Could u show, if the data is inside Postgres database ? Please

[@andfanilo]

Hello! I can't write a link in comments, but if you browse through the PandasAI documentation, at the "SQL connectors" page you'll find an example using a Postgres connector

[@mattiaselisson55]

might be a dumb questions, but since i use my API key. will there be any costs in using this method?

[@Jeevan_prakash]

No

[@sebacisterna]

Yes. But if your account is new, or you haven't used the API key before, you will spend your free credits first

[@HighTechGeeks]

if i have my own csv data how do i convert to pkl

[@andfanilo]

Hello! You can read your CSV into a Pandas file with pd.read_csv then convert to pickle with df.to_pickle
Though these days I prefer recommending .to_parquet as a binary format

[@PauloRoberto-vj4jb]

It looks like I have to be a paid user for OpenAI, right?
If so, any free alternatives that could be used with streamlit and pandas?
Thank you in advance

[@majidmohd100]

we can make use of gemini from Google

[@osamayaqoob411]

ERROR: Could not find a version that satisfies the requirement pandasai==1.4.10 (from -r requirements.txt (line 2)) (from versions: none)
ERROR: No matching distribution found for pandasai==1.4.10 (from -r requirements.txt (line 2))
hi i am facing this issue can you please resolve this issue for me ?

[@gopikrishna8850]

ModuleNotFoundError: No module named 'pandasai.callbacks' How to resolve this issue

[@andfanilo]

Hello,
Apparently pandasai removed the pandasai.callbacks module in its more recent releases, so I have pinned pandasai to use version 1.4.10, can you try downgrading?

[@Sim35may]

yes! with pip install pandasai==1.4.10 you'll fix the problem automatically

[@bangyor5949]

@@andfanilo Is there any workaround instead of downgrading the version?