NVD Cybersecurity Dataset: 337,000+ CVEs from 1988 to 2026

A repository of over 337,000 Common Vulnerabilities and Exposures (CVE) records sourced from the National Vulnerability Database (NVD). The dataset includes CVSS scores and other NVD metrics, covering vulnerabilities reported from 1988 through 2026. It is hosted on Kaggle and described as a clean collection of this security data.

Use Cases

• Trend analysis of reported vulnerabilities over time based on the 1988–2026 coverage.
• Modeling and predicting vulnerability severity based on included CVSS scores.
• Benchmarking security posture by correlating CVE metrics with other datasets.
• Training machine learning models for risk assessment using the described NVD metrics.

Strengths

• Contains over 337,000 CVE records, providing substantial volume for analysis.
• Covers a long temporal range from 1988 to 2026, enabling longitudinal studies.
• Includes structured CVSS scores, a key metric for vulnerability severity.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is precisely stated, but other metadata like file format and license are unknown.
• Last update date is unknown; freshness unverified.

Provenance

Source

National Vulnerability Database (NVD)

Collection Method

Likely aggregated and cleaned from the official NVD feeds.

Time Range

1988 to 2026

Freshness

Last updated date is unknown.

Geography

Global (software vulnerabilities are not geographically bound)