Idiomatic Pandas
summary
Five concrete Pandas practices produce big wins: enforce correct dtypes and convert wide integers to smaller numpy types (example converting columns to int8 conversion and unsigned types), convert low‑cardinality strings to category to drop memory from ~19 MB to about memory down to 6 MB, prefer chained, pipeable operations and avoid inplace mutation, avoid apply for numeric work because vectorized ops give ~40x speedup, and use groupby/agg/unstack and rolling means for fast, expressive summaries and plots.
AI-written summary · about
chapters (15)
- 0:00Introduction
- 3:40Types
- 6:35Integers
- 10:30Float
- 11:27Missing Values
- 12:15Handling Missing Values
- 13:47Memory Usage
- 15:11Convert Drive to Category
- 18:25Chaining
- 21:33Dont Mutation
- 23:04Dont Use Apply
- 25:07Exceptions
- 27:11Aggregation
- 28:57Plotting
- 30:18Multiindex
related talks
28mIdiomatic Pandas
32mPandas May Be Slow, but Pandas Doesn't Have to Be
29mMaking Pandas Fly
35mFast Pandas Performance Tricks You Wish You Knew Earlier
44mHigh Perfomance Pandas
01h 17mPandas from the Inside / "Big Pandas"
34mSmall Big Data: using NumPy and Pandas when your data...
01h 02mPandas from the inside