The past year + the future

Hi, I was just browsing the roadmap:
https://share.streamlit.io/streamlit/roadmap

It reminded me that I have been using Streamlit for about a year now and I am amazed at how much functionality has been added in that time. The past year has massively improved my own user experience (probably forms was the biggest help to me) and I am really excited looking at the roadmap for these additional features.

Thank you streamlit team for this game-changing product!

8 Likes

That is so wonderful to hear - thank you!! We are also working in the next month to add a bunch more to the roadmap to give you a much clearer view of what we’re going to be working on over the next year.

Also would love to hear more about what you’re working on and any specific features or improvements you’d love for us to prioritize for the next year.

1 Like

Ah that’s cool to hear there may be more stuff on the roadmap soon!

What I am working on:
I work for a council in New Zealand monitoring water quality in our region.
I’ve used Streamlit to build some internal data apps including:

  1. An app where you can choose your favourite swimming site and see if it is safe to swim.
  2. An app to get the raw water quality data from any site as an xlsx
  3. An app to display the results to our staff of surverys we have run with the public
    The above apps link to internal databases so can’t be hosted on the SL cloud
  4. An app using streamlit-folium where you can choose a region and I will return to you a shapefile with our water quality sites as a layer. (I hope to host on SL cloud a version of this).
  5. An image processing app which splits the image into superpixels and the user can click on the superpixels which contain algae. This is to help train our staff to identify algae in the rivers. This one could also potentially go on the cloud.
  6. Various other niche apps.

In terms of future features, improvements. The main two are
a) I am not a network guy, and some of our apps can’t be hosted on the cloud due to database access etc. We do have servers, but they run https, so I have to figure out how to host these apps on our servers. So some guides on how to do this could be useful.
b) Native https support for apps would be useful.
c) For the app 5) above I still need to figure out how to get the pixels of the returned cells when the user clicks on the image. But the rest of going well!

Thanks again for this awesome package!

1 Like

@LukeF your post is really inspiring, thank you!

In my own words: you are showing an image to the user (via st.image? or is it a chart?) and want the user to be able to click somewhere on that image and get the x/y coordinates in the Streamlit app / on the Python side? Or do you want the user to select a region of that image? Asking because we often hear such requests and I’m trying to understand the exact use cases. I don’t think we’ll work on this very soon (since it’s a tricky problem to solve!) but it’s definitely on our radar.

Hi, thanks for the reply!
Funny thing, I’m on a train right now on my way to present some of the tools we have built with streamlit to the wider sector!

At the moment I am reading an image and plotting it as a matplotlib chart.
I have not integrated this into streamlit yet, since I have been working on the Python image processing part.

For example, if you see this link (only really need to look at the images on the page)

I will segment the image to create these “superpixels” and I want the user to click all those large superpixel regions that are red, for example. When the user clicks a superpixel I will do some back end processing. For example, I may return an accuracy score (you found 8/12 red regions, for example) the show them a “ground truth” image.
I hope it is clear :slight_smile:

After I posted this yesterday I was looking around and I found the streamlit-drawable canvas component.
That looks like a possibility for me to use to achieve the desired outcome as it allows the user to click and it returns the (x,y) coordinate to Python.

1 Like

Got it, thx for the info. Cool use case and good luck with your presentation! If you find a solution with the drawable component, let me know, would be great to have a look!

1 Like