Datascience on the web
summary
Demo shows how a Jupyter workflow becomes a Flask web app: start with hello.py and inherit.py, serve on port 5000, render Jinja2 templates and handle GET/POST forms to collect square footage and condition. Dataset is ~21,000 Seattle houses; simple linear model scores CV ≈ 0.48 and a train/test run reaches ≈ 0.67. Model gets saved with joblib and metadata; geocoding uses Nominatim; plots publish via mld3 as embeddable SVG. Jupyter → Flask pipeline, Flask + Jinja2, joblib, Nominatim, mld3
AI-written summary · about
chapters (7)
- 0:00Welcome
- 2:13Flask web app basics
- 15:40Jinja2 templating
- 34:22Jupyter to web model
- 46:04Web form for input
- 1:03:20Visualizing with mpld3
- 1:12:48Live demo and deployment
related talks
45mData Science Apps: Beyond Notebooks
01h 18mBuild Data Apps by Deploying ML Models as API Services
37mBuild a Data-Driven Web App That Everyone Can Use
21mWeb apps for data science using streamlit
27mBuilding web-based engineering applications with JupyterLab components
37mPatterns for Collaboration Between Data Scientists and Software Engineers
01h 16mMake your research interactive with Jupyter Dashboards
34mTo The Web and Beyond