Using Python pandas for scientific Research
summary
Claim: pandas plus SciPy/Numpy/Matplotlib gives a practical alternative to R/Matlab for messy research data. Examples: loading CSV/Excel (hundreds of megabytes), convert messy date columns to datetime with four lines, filter and pivot and merge large tables with SQL-style joins in memory, produce reports and PDFs — generate 100 PDF tax letters in minutes. Also run K‑means on the 200‑row Swiss banknote set, get clear clusters with a few misclassified genuines, demonstrating workable analysis and big time savings.
AI-written summary · about
chapters (23)
- 0:00Introduction
- 0:50Agenda
- 1:28About Uwe
- 2:58Python
- 4:35Hello World Example
- 4:57What is pandas
- 6:45Python framework
- 8:24Python distribution
- 9:41Data handling
- 9:57Data structures
- 11:39Data frames
- 15:36Open CSV file
- 17:38Python code
- 18:03Selection and filtering
- 21:39Merge data
- 23:33Transform data
- 27:38Check payment
- 28:32Questions
- 34:36Boxplot
- 35:17Scatterplot
- 35:54Cluster Analysis
- 40:17Conclusion
- 41:13Literature Recommendations
related talks
44mIntroduction to Pandas
01h 16mIntro to Data Analysis with Python Data-table
01h 24mIntroduction to Data-Analysis with Pandas
36mAutomating Data Analysis with Pandas
45mPython useRs
01h 04mVisualising Data through Pandas
14mKeynote: Scalable Python We Can Afford
01h 31mPandas, Data Wrangling & Data Science