Revised 19 October 2026 · v1.0.0 · what changed

Code: github.com/herrrickshaw/kalki_flooding

Watch the walkthrough

Watch on YouTube: youtu.be/JcUzIvraEho

Why it is useful

• Urban flood questions need long rain records; the repo ships city daily series so you can recompute extreme-rain metrics without downloading raw grids.
• It uses IMD's official intensity categories (heavy 64.5 mm, very heavy 115.6 mm, extremely heavy 204.5 mm), so results map to terms in weather bulletins.
• A DATA_CAVEATS file documents where gridded data misleads (Pune cell, sea-masked Mumbai and Chennai points); this matters before citing any number.
• Evidence from the run: city_rain_metrics.py regenerated 1,623 city-years in about 3 seconds and the output CSV was identical to the committed one.

What it is for

Recompute per-city yearly rainfall metrics (annual, monsoon, rainy days, heavy-day counts, wettest day) and charts from the committed daily CSVs.

The repo also holds collector.py and request_access.py for the IUDX flood-sensor network. Those need registered client credentials and provider approval and were not run.

Who it is for

• Students, journalists and planners exploring long-run heavy-rain trends
• Analysts who want a documented starting dataset for city flood-risk work

What you need

Get the code and an AI assistant

You do not have to be a programmer. Get the repository, give it to an AI assistant, and ask in plain language. Any of these routes works:

git clone https://github.com/herrrickshaw/kalki_flooding.git
cd kalki_flooding
claude      # or: gemini

Install

1. Clone and create a virtualenv.

   git clone https://github.com/herrrickshaw/kalki_flooding.git
   cd kalki_flooding/analysis
   python3 -m venv .venv && .venv/bin/pip install matplotlib

Quick start: commands and what you should see

Rebuild the metrics and two charts

.venv/bin/python city_rain_metrics.py

Prints 'wrote .../city_rain_metrics.csv (1623 city-years)' and 'wrote city_monsoon_trend.png, city_extreme_days.png'. It overwrites those files in the folder, so run it in a copy if you want to keep originals.

Sample outputs

Real output from a run on the date shown in the sources line; trimmed to fit.

Rebuild run

.venv/bin/python city_rain_metrics.py
wrote .../city_rain_metrics.csv (1623 city-years)
wrote city_monsoon_trend.png, city_extreme_days.png

Output CSV matched the committed file.

Heavy-rain days per year, early vs recent, from the CSV (Pune cell excluded)

pandas: mean heavy_days for years <=1960 vs >=1996, per city
city       1901-60  1996-2025  wettest day (mm, year)
ahmedabad    1.88     2.47      384.3 (1927)
bengaluru    0.82     1.20      188.5 (1998)
chennai      3.40     3.20      329.3 (2015)
delhi        1.20     0.53      234.6 (1963)
kolkata      1.82     2.80      326.2 (1978)
mumbai      12.42    13.30      423.4 (2005)
surat        3.42     5.23      400.4 (1941)

Gridded 25 km cells mute extremes (Delhi, Mumbai); see DATA_CAVEATS.md. Differences are descriptive, not tested for significance.

Ways to use it

Sample prompts for your AI assistant

Copy, paste and edit the parts in [brackets].

• Read analysis/DATA_CAVEATS.md and summarise which cities' numbers I should not compare.
• Using city_rain_metrics.csv, plot heavy_days by decade for Surat and Kolkata.
• Explain the IMD intensity categories used in city_rain_metrics.py.
• Suggest a significance test for the early vs recent heavy-day difference.
• Explain what collector.py does and what I must do before running it.

Limits and honest caveats

Sources: the kalki_flooding README, analysis/DATA_CAVEATS.md and committed city_daily CSVs (IMD gridded data); city_rain_metrics.py run on 19 October 2026 in a scratch copy with a fresh virtualenv. Other scripts and the IUDX collector not run.

About this article: Researched, written and edited by Umashankar Triplicane Dwarakanathan, with AI research assistance; every figure is meant to trace to the primary source cited. See the Editorial Policy for how sourcing, AI use and corrections work.