Using Python in a Data Hackathon
summary
Shows a practical weekend workflow: prebuilt Anaconda/Jupyter shell scripts to avoid setup fights, use SQLite for portable ingestion, and prefer binary interchange like feather over CSV. Recommends fast baseline modeling with XGBoost, then focus on feature engineering and simple ensembling; a one-line business rule gives a 94% baseline and the team reaches 96%, finishing 27th. Mentions alternatives tried (Dask, TPOT, Keras) and warns against cloud complexity and fancy object DBs.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 1:45The problem
- 2:18What do you do
- 3:17Overview
- 4:11Data Formats
- 5:40Data Types
- 6:26Environment Setup
- 12:55Data Processing
- 16:12Hints Tips
- 17:32Results
- 18:25Feature importance
- 19:47Final solution
- 22:05Questions
related talks
30mEmpowered by Python - A success story
41mKeynote: Making the Big Data ecosystem work together with Python
34mPython as a Framework for Analytics and Growth Hacking
35mWorking with data in Python for the Aspiring Data Scientists
31mData Science Project for Beginners
35mMental Models to Use and Avoid as a Data Scientist
45mPython useRs
01h 25mAdvanced Excel Analytics: Python Integration Workshop