When Cold Broke More Records Than Heat

Executive Overview

Climate narratives often focus on rising heat. But when 2025 temperature data for Barquisimeto was compared against a 10-year baseline (2015–2024), a surprising pattern emerged: cold extremes outpaced heat extremes. Out of 52 record-breaking days identified, 28 broke cold records and only 24 broke heat records — suggesting that the real story is not a simple warming trend, but increased temperature volatility.

This project was built as part of an Advanced Data Science with Python course, using real NASA Power daily temperature data and a reproducible analytical pipeline.

Key Findings

• 52 record-breaking days in 2025 — out of 365 calendar days analyzed.
• 28 days broke cold records vs. 24 that broke heat records — more extreme cold events than extreme heat, a counterintuitive result for a semi-arid city.
• Increased volatility, not just warming: the data points to wider temperature swings rather than a single directional climate shift.
• Reproducible methodology: the full pipeline — from raw NASA data to published visualization — is available as an open Jupyter Notebook.

Temperature Record Visualization

The chart below shows the 10-year baseline temperature envelope (record highs and lows for each day of the year), with 2025 record-breaking days overlaid as scatter points. Red dots indicate days where 2025 exceeded the historical high; blue dots where it fell below the historical low.

Daily temperature envelope for Barquisimeto (2015–2024 baseline) with 2025 anomalies highlighted. Data source: NASA Power.

Analytical Approach

Data Source: NASA Power Project — daily surface temperature data (T_max and T_min) for Barquisimeto, Venezuela (2015–2025).

Methodology:

• Loaded and cleaned 10 years of daily temperature records, removing leap-day (Feb 29) for consistent 365-day comparison.
• Computed record highs and record lows for each calendar day across the 2015–2024 baseline period.
• Identified 2025 days where the daily high exceeded the historical record high, or the daily low fell below the historical record low.
• Visualized results as a temperature envelope chart with anomaly scatter points, exported as PNG for dashboard embedding.

Tools: Python (pandas, matplotlib, folium) · Jupyter Notebook · NASA Power API · GitHub Pages