Tucson-vanishing-winters

Tucson’s Vanishing Winters — Watch the Blues Fade and Reds Take Over

University of Arizona Data Visualization Challenge 2026

Submitted by Vanishree Venkatachalapathy · March 18, 2026

🌐 View Full Report (Live Site)     ✨ Open Interactive Particle Field

About

Tucson, Arizona sits in the Sonoran Desert — a city where summer heat is expected, but where mild winters have long defined its character. This project asks a simple question: has that changed?

Using 41 years of monthly NASA satellite temperature data (1984–2024), this visualization tracks how Tucson’s thermal fingerprint has shifted across every month and every year of the record. Rather than a single annual average, each month is measured against its own 41-year baseline — so a warm January is judged on January’s terms, not July’s.

The answer is clear: the blues are fading. The reds are here to stay.


Visualizations

Five complementary views of the same 41-year record:

Visualization What It Shows
Warming Stripes One stripe per year, colored by annual mean anomaly — no axes, just the pattern
Annual Anomaly Trend Year-by-year anomaly bars with a LOESS smoother revealing the long-term trajectory
Monthly Heat Calendar All 492 month–year cells as a color grid — spot which months and seasons are shifting
Monthly Warming Rose Polar chart showing which calendar months have warmed most relative to their own baselines
Generative Particle Field An animated, interactive canvas encoding all 492 cells simultaneously as drifting particles

📊 View Full Report with All Visualizations — all five charts in one Quarto document

Open Standalone Interactive Particle Field — full-screen animated visualization

The Interactive Particle Field

The centerpiece visualization maps every (month, year) pair onto an animated grid of particles:


How to View

No installation required. Choose one of the options below:

Option Link
🌐 Full report (Quarto document + all charts) Live Site
✨ Standalone interactive particle field Open Interactive
💾 Offline Download tucson-vanishing-winters.html from the code folder of this repo and open in any modern browser

How to Reproduce

To render the document from source:

Requirements

Install R packages

install.packages(c("nasapower", "dplyr", "tidyr", "lubridate", "jsonlite",
                   "htmltools", "ggplot2", "ggtext", "glue", "knitr", "scales", "here"))

Render

Run the following from the project root directory (tucson-vanishing-winters/):

quarto render code/tucson-vanishing-winters.qmd

The raw data is pre-downloaded and stored in data/tucson_T2M_monthly_1984-2024.csv. No API call or internet connection is required.

Expected output

After rendering, two files are created:

Open either file in any modern browser.


Data

Field Detail
Source NASA POWER Agroclimatology API
Variable T2M — Monthly mean 2-metre air temperature (°C)
Location Tucson, Arizona · 32.22°N, 110.97°W
Period January 1984 – December 2024
Observations 492 (41 years × 12 months)

Methodology

Temperature Anomalies

Temperature anomalies are calculated on a per-month basis: each observation is compared to that specific month’s 41-year mean, not a single annual average. This preserves the seasonal structure of the warming signal — a warm January is judged against January’s own average, ensuring no month is penalized or rewarded by comparison to a different season.

Generative Particle Field

The grid is divided into 41 columns (one per year) and 12 rows (one per month), creating 492 cells. Each cell is assigned hundreds of small moving dots — particles — that drift around their home cell in slow, flowing motion. The color of each particle reflects the temperature anomaly of its cell: red for warmer than average, blue for cooler, and amber for warm summer months.

The flowing movement is generated using a hand-coded fractional Brownian motion (fBm) noise algorithm — no external library. Think of it like gentle wind gusts that nudge each particle in a slightly different direction every frame, while an invisible elastic force keeps pulling it back toward its home cell. The result is a visualization that feels alive while remaining anchored to the underlying data.


Tech Stack

Layer Tools
Data retrieval R · nasapower
Data wrangling R · dplyr, tidyr, lubridate
Static visualization R · ggplot2, ggtext, scales, glue
Interactive visualization Vanilla JavaScript · Canvas 2D API
Noise algorithm Hand-coded fractional Brownian motion (fBm) · no external library
R → JS bridge jsonlite (JSON export) · htmltools (iframe embed)
Document Quarto HTML · self-contained single file

Repository Structure

tucson-vanishing-winters/
├── README.md
├── LICENSE
├── .gitignore
├── data/
│   └── tucson_T2M_monthly_1984-2024.csv     
├── code/
│   ├── tucson-vanishing-winters.qmd          
│   └── tucson-vanishing-winters.html          
└── output/
    └── tucson-vanishing-winters-interactive.html


Citation

Data: NASA POWER Project. (2024). Prediction of Worldwide Energy Resources (POWER) — Agroclimatology Climatology Dataset. NASA Langley Research Center. Retrieved from https://power.larc.nasa.gov/

This project: Venkatachalapathy, V. (2026). Tucson’s Vanishing Winters: Watch the Blues Fade and Reds Take Over. University of Arizona Data Visualization Challenge 2026.


Author

Vanishree Venkatachalapathy University of Arizona · Data Visualization Challenge 2026


License

This project is licensed under the MIT License — see LICENSE for details.


Acknowledgments

Temperature data provided by the NASA POWER Project, funded by the NASA Earth Science Directorate Applied Science Program.