Lessons learned from applying PyData to our marketing organization
summary
GetYourGuide describes a PyData-stack system that automates keyword mining, campaign automation, predictive modeling, omni‑channel integration and customer segmentation. The stack uses Celery + RabbitMQ for API rate control, Fabric for deploys, Alembic-tracked schemas (≈250 schema versions in ~600 days), IPython notebooks as the operator UI, and a mix of scraping and third‑party APIs for ranking data. The pipeline automates 75% of work, launches campaigns in 13 markets and scales marketing 3x. Many components succeed; overly complex keyword NLP is abandoned for template‑based, pragmatic solutions.
AI-written summary · about
chapters (16)
- 0:00PyData in marketing
- 1:55Marketing intelligence with PyData
- 3:22Lessons from domain expertise
- 5:07Speed in marketing competition
- 6:52Marketing technology components
- 10:46Building marketing infrastructure
- 13:56Data storage and schema evolution
- 15:23Third-party API challenges
- 17:28Data scraping risks
- 18:55Analyzing marketing data
- 21:41Market size and competitor analysis
- 23:51Sentiment analysis of reviews
- 26:40Keyword and ad targeting
- 27:45Hiring for marketing data science
- 29:10Measuring prediction impact
- 30:13Summary
related talks
40mPyData Stack: Pure Python open source data platforms
31mBuilding Data-Driven Client Relationship Management in Banking with Python
39mHow Python Drives Modern Data Engineering
48mThe Use of Python in the Data World
30mEmpowered by Python - A success story
28mUsing Python in a Data Hackathon
21mSaaS Compliance Automation with a Python Stack
46mThe Python Data Ecosystem: Navigating a fragmented landscape