Tech Tutorial: Create Twitter and APIs
summary
Builds an XGBoost regression with tidymodels on Inside Airbnb listings for Vienna, tunes via 5‑fold CV and a grid (best found mtry=5, ~2000 trees), logs price and uses a recipe for preprocessing; model flags underpriced listings (example: $11 listing vs $56 prediction → ~80% discount) and surfaces the top 500 candidates on a Leaflet map. Model and report deploy as a Quantargo API, and a bot using rtweet and meme posts visualizations twice daily from a scheduled workspace.
AI-written summary · about
chapters (20)
- 0:00Start
- 1:56Agenda
- 2:18The 30 Day Sustainability Data Challenge
- 4:48Introduction
- 8:16Markdown Document
- 18:09Ggplot
- 31:55The 30-Day Sustainability Data Challenge
- 32:2030 Days Sustainability Data Challenge
- 35:19Twitter Bot
- 41:31Create a New Twitter Bot
- 1:02:15Documentation
- 1:03:26Find Undervalued Airbnb Apartments in Vienna
- 1:06:02Listings Data Set
- 1:19:34Modeling
- 1:22:24Create a Recipe
- 1:38:56Create the Model
- 1:42:10Final Tuning Procedure
- 1:42:50Cross Validation Data Set
- 1:44:26Tuning Grid
- 2:00:25Collect the Metrics
related talks
57mGood Practices for Applied Machine Learning
22mModelling, Tidyverse Way- 57mGood Practices for Applied Machine Learning
23mTotal Tidy Tuning Techniques
43mDeveloper Accessible Machine Learning
18mtidymodels: Adventures in Rewriting a Modeling Pipeline
01h 02mA developers guide to Machine Learning
24mChallenges in building Machine Learning models in production