Data Science · Chapter 9 of 43
Reading & Writing Data
pandas can read/write CSV, Excel, JSON, Parquet, SQL and more with one-line functions.
Parquet is preferred for large tabular data — it's columnar, compressed and preserves types.
Example 1 (python)
import pandas as pd
df = pd.read_csv('data.csv')
df.to_parquet('data.parquet')CSV in, Parquet out.
Example 2 (python)
df.to_excel('report.xlsx', index=False)
df2 = pd.read_json('logs.json')Excel and JSON I/O.
Key points
- CSV: universal, human-readable.
- Excel: for business stakeholders.
- Parquet: fast + typed for large data.
- index=False avoids extra columns on export.
💡 Note: For big data, prefer Parquet over CSV — same data, often 5-10× smaller and much faster to load.