How Dimensional is a `pandas.DataFrame`, anyway?
summary
Claim: a pandas.DataFrame is one-dimensional, an index-aligned collection of Series, not a true 2D array. Demonstrations show Series as strictly homogeneous columns, DataFrame = indexed collection of 1‑D Series, xarray pricing tables misrepresent sparse time series, and stack/unstack acts like turning one homogeneous dataset into many. Empirical failures include DF.describe() coercing dtypes and losing precision, groupby across columns forcing transposes and dtype churn, and a trading model rebuilt as a 1D multi‑index that actually works.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 1:05Pandas is one-dimensional
- 3:31Homogeneity in Python data
- 10:10Python dictionaries dual nature
- 11:13Transition to NumPy and lists
- 13:58Fixed vs dynamic size arrays
- 15:22Dimensionality of nested lists
- 17:06Pandas coordinates: structural vs data
- 19:12Series and DataFrame as 1D
- 20:57Pandas API mistakes from dimensions
- 21:19Xarray 2D data coercion
- 22:42Pandas stack and unstack
- 24:07Trading data with 1D series
- 27:55Correct trading data modeling
- 28:36Summary
- 28:58Q&A: Group by axis deprecation
- 30:21Q&A: Relational data model
- 31:48Q&A: Pandas usability
related talks
31mEffective Data Analysis with Pandas Indexes
01h 02mPandas from the inside
01h 17mPandas from the Inside / "Big Pandas"
41mWe don't dataframe shame: A love letter to dataframes
44mIntroduction to Pandas
38mMind the Gap! Bridging the pandas – scikit learn dtype divide
35mIdiomatic Pandas
35mA Tour of the Many DataFrame Frameworks