Southwest heatwave city temperatures

This is the minimal script a human user would actually run: load data, call EchoTime, inspect the returned object.

from pathlib import Path

import pandas as pd
from echotime import compare_series, profile_dataset

data_path = Path(__file__).resolve().parents[1] / "data" / "real_heatwave_city_temps_2024.csv"
df = pd.read_csv(data_path)

report = compare_series(df["phoenix_temp_max"], df["las_vegas_temp_max"], left_name="Phoenix max temp", right_name="Las Vegas max temp")
profile = profile_dataset(df[["phoenix_temp_max", "las_vegas_temp_max"]], domain="climate")

print(report.to_summary_card_markdown())
print(profile.to_summary_card_markdown())

You should see high pairwise correlation metrics plus a dataset profile explaining the broader panel structure.

# EchoTime similarity summary

**Compared:** Phoenix max temp vs Las Vegas max temp

## Headline

Phoenix max temp vs Las Vegas max temp: Pearson r 0.73, Spearman rho 0.75, Kendall tau 0.56. The best agreement appears in trend similarity. The weakest agreement appears in shape similarity and derivative similarity, so timing or regime differences probably matter.

## Familiar statistics

| metric | value |
|---|---:|
| Pearson r | 0.727 |
| Spearman rho | 0.746 |
| Kendall tau | 0.556 |
| Best-lag Pearson r | 0.822 |
| Mutual info | 0.281 |
| First-difference r | 0.123 |

## Time-series-specific metrics

| plain-language label | score |
|---|---:|
| trend similarity | 0.915 |
| spectral similarity | 0.605 |
| dtw similarity | 0.586 |
| shape similarity | 0.299 |

## Recommended next actions

- Plot both series after z-score normalization to show the shared shape without scale differences.
- Run rolling or windowed similarity if you expect the relationship to change over time.
- Use structural-profile similarity when scales, frequencies, or observation modes differ too much for raw-shape comparison.

• Pearson 0.73 / Spearman 0.75 / Mutual info 0.28 shows the two cities track each other closely through the heatwave window.
• The supporting dataset profile shows the broader structural context behind that pairwise comparison.
• This is a realistic wide-table workflow you can reuse for any small panel of real measurements.