FRED NASDAQ Composite (Mirror)
- Rows
- 13,960
- Columns
- 2
- Data range
- 1971-02-05 to 2026-06-18
Direct CSV mirror of FRED NASDAQCOM daily index. Time-series stock data.
AI analysis
FRED NASDAQ Composite (Mirror) — Dataset Analysis
1. Dataset Overview & Correlation Value This dataset is a daily time-series of the NASDAQ Composite closing index, mirrored directly from the Federal Reserve Economic Data (FRED) system via a CSV feed hosted at eco3min.fr. The source URL suggests a third-party mirror rather than a direct FRED pull, which warrants attention regarding freshness and update cadence — though the declared time coverage extending to 2026-06-18 implies the mirror is actively maintained and potentially current. With nearly 54 years of daily observations (1971–2026), this is an exceptionally long financial time-series, making it highly valuable for long-horizon correlation studies involving macroeconomic cycles, interest rate regimes, inflation periods, and technology sector growth arcs.
2. Data Quality Observations Data quality is notably strong for a financial time-series. Zero null cells across both columns eliminates any imputation concerns, and the 13,960 distinct dates matching the total row count exactly confirms there are no duplicate date entries — a critical integrity check for time-series joins. The nasdaqclose column shows 13,432 distinct values out of 13,960 rows, meaning roughly 528 rows share a repeated closing value with another row — this is expected behavior in markets (identical closes do occur), not a data flaw. No type mismatches are reported. The one caution is the third-party mirror provenance from eco3min.fr; downstream users should periodically validate against the canonical FRED NASDAQCOM series, particularly for the most recent observations.
3. Key Column Distributions The nasdaqclose column tells a vivid story of exponential market growth. The mean of 3,295.54 sits dramatically above the median of 1,511.14, immediately signaling strong right skew — confirmed by a skewness coefficient of 2.28. The interquartile range spans 287.19 (Q1) to 3,443.32 (Q3), reflecting that half the dataset's history occurred when the index was below ~3,443, while the maximum of 27,093.9 captures the 2020s bull run. The standard deviation of 4,934.19 — larger than the mean itself — underscores how non-stationary this series is in raw level form. The 1,665 flagged outliers (~11.9% of rows) are almost certainly concentrated in the post-2020 period and are not data errors but structural regime shifts; any correlation analysis should work on log-returns or percentage changes rather than raw index levels to stabilize variance and remove the trend.
4. Recommended Join Key The date column is the unambiguous join key — it is a clean Date type, has zero nulls, and is perfectly unique (13,960 distinct values across 13,960 rows). For cross-dataset joins, standardize to ISO 8601 date format (YYYY-MM-DD) and account for trading day gaps (weekends, US market holidays) when joining against daily macroeconomic series that may publish on all calendar days. A calendar or trading-day lookup table may be needed as an intermediary for datasets with different publication cadences (e.g., monthly CPI, weekly jobless claims).
5. Recommended Pairing Datasets Several dataset categories would yield high-value correlations with this series. Federal Funds Rate or Treasury Yield data (also available via FRED) would illuminate the well-documented inverse relationship between interest rates and equity valuations. S&P 500 or Dow Jones daily close data would allow index divergence analysis, particularly to isolate tech-sector-specific dynamics. US CPI or PCE inflation series could reveal how inflationary regimes correlate with NASDAQ performance over multi-decade horizons. VIX (Volatility Index) data would pair naturally for risk-sentiment studies. Finally, Big Tech earnings or revenue time-series (FAANG constituents) could help explain post-2010 index concentration effects, since the NASDAQ Composite is increasingly driven by a small number of mega-cap names.
Columns
- date (date)
- nasdaq_close (decimal)