Streamlit
Welcome, champion! 🔥
You've mastered 0 of 81 lessons. Start your first lesson free — no signup needed.
About this course
A Streamlit app looks simple until a button forgets it was clicked, a slow query runs again on every slider move, or a change made for one user shows up for everyone. Almost all of these problems come from one idea: the whole script reruns from top to bottom on every interaction. This online Streamlit course on Abekus starts from that rerun model and builds up to widgets and forms, layout and multipage navigation, session state, caching, charts and dataframes, connections and secrets, testing and deployment, in short reading-card lessons each followed by an MCQ quiz.
Who this Streamlit course is for
It is written for Python users who want to turn a script into a web app that others can use, without doing front-end work: analysts, data scientists and ML engineers. You should already write Python and be able to read pandas DataFrame code; the course does not teach pandas, plotting libraries or LLM APIs. By the end you can build multipage apps with widgets, forms and session state, cache data and resources correctly, control reruns with fragments, add sign-in and secrets, test an app headlessly and deploy it.
Examples follow one modern setup: Streamlit 1.67 on Python 3.11 or later, started with streamlit run app.py, where every example is one small script. Data is held in pandas DataFrames, charts use the built-in chart functions and Altair, multipage apps are defined with st.navigation and st.Page, SQL access goes through st.connection with SQLAlchemy, secrets live in .streamlit/secrets.toml and tests use AppTest under pytest. Removed APIs such as st.cache and the st.experimental functions are named only so that you can recognise and replace them. The course leaves out building custom components in React, and the account and settings screens of Community Cloud and Snowflake. It is an independent course and is not affiliated with or endorsed by Snowflake or the Streamlit project.
How the lessons work
Each lesson is a short set of reading cards on one function or behaviour, usually with a script of a few lines and a description of what the app shows, followed by a quiz of multiple-choice questions that often ask what happens on the next rerun. Lessons go from basics up: a first app with a slider and a chart, then widgets and input, layout and navigation, state and reruns, caching, data and charts, connections, secrets and users, configuration, testing and deployment, advanced topics such as chat apps and async code, and finally comparisons and worked cases. You move at your own pace.
MCQ practice vs video courses for Streamlit
Video courses on Udemy, Coursera or YouTube are a good choice if you want to watch a dashboard being built in an editor from an empty file. This course is built for a different job: a short explanation followed straight away by questions about what a script will do, such as whether a variable survives a rerun, what a button returns on the run after the click, when a callback fires, whether two users share a cached object, or which part of the page a fragment reruns. Streamlit's execution model is precise, and predicting it question by question builds the mental model that saves debugging time later. Many learners use both: a video for a first walkthrough, and questions to make the details stick.
Common Streamlit mistakes
The course corrects these directly:
- Expecting a button to stay pressed. st.button returns True only for the rerun caused by the click; session state is what remembers it.
- Treating ordinary variables as memory. Every interaction reruns the script from the top, so plain variables start fresh each time.
- Caching a database connection with st.cache_data. Connections and models belong in st.cache_resource or st.connection.
- Changing an object returned by st.cache_resource. It is one shared object, so the change is seen by every session.
- Assuming session state is shared between users. It belongs to one session, which is one browser tab.
- Copying removed APIs from old tutorials. st.cache and st.experimental_fragment are gone; st.cache_data, st.cache_resource and st.fragment replace them.
- Committing secrets.toml to git. Secrets are read through st.secrets, and the file stays out of the repository.
Best way to learn Streamlit
Learn the rerun model first and trace every script by hand: what runs on load, what runs after a click, and what is kept between the two. Then add one tool at a time for the problem it solves: session state to remember, forms to batch input, caching to avoid repeated work, fragments to rerun less of the page. The last part of the course is built around these choices, with cache_data versus cache_resource, session state versus caching, when to use forms or fragments and when Streamlit is the wrong tool, applied to a sales KPI dashboard, a CSV cleaning tool for analysts, an LLM chat app and an ML model demo app. Comparisons with Dash and Flask show where each one fits.