Data Ingestion with Python

Learn how to use Python tools and techniques to solve one of the main challenges data scientists face: getting good data to train their algorithms.

Introduction

• Why is data ingestion important?
• What you should know
• Using the exercise files
• Using the Coderpad quizzes

1. Data Ingestion Overview

• Overview of data scientists work
• Where does data come from?
• Different types of data
• The data pipeline (ETL)
• Final destination (data lake)

2. Reading Files

• Working in CSV
• Working in XML
• Working in Parquet, Avro, and ORC
• Unstructured text
• JSON
• Solution: CSV to JSON

3. Calling APIs

• Working with JSON
• Making HTTP calls
• Processing event-based data
• Solution: Location from IP

4. Web Scraping

• Try to find an API
• Working with Beautiful Soup
• Working with Scrapy
• Working with Selenium
• Other considerations
• Solution: Get stock information from HTML

5. Schema

• What are schemas?
• Working with ontologies
• What should be in schema
• Schema changes
• Schema validations

6. Working with Databases

• Types of databases
• Hosted and cost of ops
• Working with relational databases
• Working with key or value databases
• Working with document databases
• Working with graph databases
• Solution: ETL

7. Troubleshooting Data

• Data is never 100% okay
• Causes of errors
• Filling missing values
• Finding outliers (manual)
• Finding outliers (ML)
• Solution: Clean rides dataset

8. Data KPIs and Process

• Design your data
• KPIs
• What to monitor?

Conclusion

• Next steps