Building large scale data apps with Taipy | Head-to-Head against Streamlit

[@DataSciencewithHarshit]

So, what do you think about Taipy?

[@narottamsharma4720]

Amazing

[@absar66]

Business License fee is expensive compared to available alternatives

[@JayaDattasaiKala]

Theres a lot of room for custom widgets to be developed, this will take a lot for taipy to compete with streamlit ; generally streamlited apps gets transformed into either reactified or any popular frontend fw with django or fast api as backend for becoming prod ready I have explored many class A frameworks from python, none of them provides a excel /table copy paste directly into browser dataframe editorwidget other than streamlit.

[@maruthiprithivi]

Very nicely structured content. Loved that you kept it to the point and didn't add fillers to drag out the duration of the video. High production quality!!

[@kubectlgetpo]

Streamlit is not designed for production. It is exactly fir prototyping, fail fast, iterate fast. Once it sticks, go proper full stack!

[@GamingCritique1]

Nice! This makes it so much easier

[@coding_wiz]

will try it for one of my dashboards. Very nice tutorial.

[@pranaamerta8015]

How's this compared to Shiny for Python

[@vishwastyagi8390]

Great ..we’ll try this Thanks

[@AlonAvramson]

Looks great ! thank you!

[@interspacer4277]

What about vs. Anvil?

[@gagansingh3481]

I love streamlit easy to understand without any such hard coding required , client needs output Designer app se koi Lena Dena Hota use

[@DataSciencewithHarshit]

Agreed, I also love Streamlit. In fact, comparing it with Streamlit does say a lot about Streamlit. But if I think beyond those use cases, that’s when I will look for an alternative.

[@irapk79]

NIce, very nice , CRUD possible ?

[@DataSciencewithHarshit]

Yes, check their gallery for more examples. You will have to hitch it up with a DB for CRUD.

[@irapk79]

@@DataSciencewithHarshit Awesome ,thank you

[@AliAlias]

Streamlit more easy to write code

[@MohanKumar-gj9th]

Streamlit is free though

[@AhmedIsam]

There is a shameless cheating @ 9:32 what the hell is Decimator object that reduces number of points from 1 million to 500 !!! Why there is no equivalent decimation in streamlit?

[@DataSciencewithHarshit]

This is already explained in the video, watch till the end. The whole point is the support for such algos within Taipy.

[@AhmedIsam]

@@DataSciencewithHarshit Thanks for the reply, but that decimation support is literally a one-liner. That has nothing to do with web framework. Here is the equivalent cheat in streamlit: ```python df = df.iloc[::len(df)//500][:500] # Decimation ``` I tried it out, streamlit is reduced to 1 second delay, still not as fast as taipy. You can still prove your point while being fair, don't make streamlit plot 1 million points and decimate the task to 500 points in taipy.

[@AhmedIsam]

@DataSciencewithHarshit I think it just clicked why Decimator object is a thing. My one-line python equivalent still means that there is round-trip to the backend to resove the reaction to user input. While the decimator object wraps a javascript solution that is front-end resolved, so, no frontend-backend-frontend trip is involved, making it faster. I will consider this solution for data intensive apps.

[@KevinKreger]

But flask is not for production.

[@DataSciencewithHarshit]

Have you checked which all companies use Flask in production?

[@iHydrogen]

what is it for then