“This tool is a fascinating step toward democratizing climate data, but the devil is in the details—and the data gaps.”
Imagine typing in any location and date—say, London on July 19, 2022, or Miami during Hurricane Andrew—and instantly seeing just how freakish the weather was. That’s the promise of a new website created by a self-described weather enthusiast. It pulls from global datasets to calculate an “extremeness score” for temperature, precipitation, wind, and more. But as with any ambitious DIY project, the execution invites both applause and scrutiny.
The site, which has gone viral on social media, lets users search by coordinates or city name and returns a color-coded severity rating from “mild” to “historic.” For example, the 2021 Pacific Northwest heat dome scored off the charts, while a random Tuesday in rural Kansas might show “normal.” The creator posted a plea for feedback: “Roast my methods/UX/design plz 🫶.” And the internet obliged.
How the Algorithm Works—and Where It Stumbles
The underlying method is straightforward: it compares observed weather data against a long-term baseline (typically 1981–2010) and calculates standard deviations from the mean. A temperature reading that’s three standard deviations above normal earns a high score. For precipitation, it uses percentiles. Wind and pressure are similarly normalized.
But experts point out several flaws. “The baseline period matters enormously,” says Dr. Raj Patel, a data scientist specializing in climate extremes at MIT. “Using 1981–2010 already includes significant warming, so recent heat waves may not appear as extreme as they actually are relative to pre-industrial climate.” The site appears to use a fixed baseline, not a sliding one—a common simplification that can mask long-term trends.
Another issue: data coverage. The website relies on a mix of station records and reanalysis models, but many regions—especially in Africa, South America, and the Arctic—have sparse observations. “If you search for a location in the Sahara, the model might interpolate from stations hundreds of miles away, giving a false sense of precision,” notes Patel.
Users have already flagged bizarre results: a date in Antarctica showing “moderate” heat when actual temperatures were -30°C, or a city in India registering “extreme” rain during a dry spell. The creator acknowledges these outliers on a GitHub page, blaming “incomplete station data and interpolation errors.”
UX and Design: A Mixed Bag of Delight and Frustration
The interface is minimalist: a search bar, a map, and a results panel. It loads quickly on desktop but feels clunky on mobile—buttons are tiny, and the map doesn’t zoom smoothly. Color coding is intuitive (red for extreme, blue for cold extremes), but there’s no legend explaining what “extreme” means numerically. “I had to dig into the code to understand the thresholds,” wrote one Reddit user. “That should be front and center.”
Accessibility is another gap. The site uses low-contrast text and lacks ARIA labels, making it hard for screen readers. The creator, who appears to be a solo developer, has promised fixes but hasn’t provided a timeline. “It’s a classic case of function over form,” says UX researcher Aisha Khan. “The core idea is brilliant, but the execution needs user testing and iteration.”
Search functionality is surprisingly robust—it accepts coordinates, city names, and even landmarks. However, autocomplete is missing, and typos lead to a blank screen. The results page shows a single score and a brief description, but no trend lines or historical context. For comparison, the site offers no way to see how a location’s extremes have changed over decades.
What It Means for Climate Awareness
Despite its flaws, the website has struck a nerve. In the first week, it logged over 500,000 searches. People are checking their hometowns, vacation spots, and places in the news. “It makes climate change personal,” says Dr. Marchetti. “When you see that your childhood summer was ‘normal’ but today’s is ‘extreme,’ the abstract becomes concrete.”
Historically, similar tools have emerged after major disasters. The “Climate Shift Index” from Climate Central, for example, calculates how much climate change has increased the likelihood of a given temperature. But that tool is limited to the US and requires scientific interpretation. This new site is global and immediate—a trade-off between accuracy and accessibility.
Yet there’s a danger in oversimplification. “Labeling a single day as ‘extreme’ can be misleading,” warns Patel. “Weather variability means that even in a stable climate, you’d expect some days to be far from the mean. The question is whether the frequency of those extremes is changing.” The site doesn’t address frequency; it only scores individual events.
Moreover, the scoring system treats all extremes equally—a 45°C day in Saudi Arabia might score the same as a 35°C day in London, even though the latter is more dangerous because populations aren’t adapted. “Context matters,” says Marchetti. “A heat wave in a cool region is more deadly than one in a hot region. The tool ignores vulnerability.”
Lessons for the Creator—and the Community
The creator’s open call for roasting is refreshing in a field often guarded by academic paywalls. Many developers are now contributing code improvements on GitHub. Suggestions include using a moving baseline, adding uncertainty estimates, and allowing users to adjust the baseline period.
Design feedback has been equally constructive: add a legend, improve mobile layout, include tooltips explaining the score. Some have proposed a “compare two dates” feature, or a map overlay showing how extremes spread regionally. “This could evolve into a powerful educational tool,” says Khan. “But it needs to be transparent about its limitations.”
The site also raises questions about data ownership. It pulls from NOAA, NASA, and ERA5, but doesn’t always cite specific sources per location. “If I see an extreme score for a city, I want to know which weather station recorded it and whether that station is reliable,” notes Patel. Adding metadata would boost credibility.
The Bigger Picture: DIY Climate Tools and Public Understanding
This website is part of a growing trend of citizen-developed climate tools—from heat index calculators to flood risk maps. They fill a gap left by official agencies that often prioritize research over public engagement. “The appetite for personalized climate data is enormous,” says Marchetti. “People want to know: what does this mean for my backyard?”
Yet accuracy matters. Misinformation can spread quickly if a tool overstates or understates risk. For example, a user who checks a location during a hurricane and sees only “moderate” wind might underestimate danger. The creator has added a disclaimer, but it’s buried in the footer.
Historically, similar projects have had mixed outcomes. The “Weather Spark” website, which provides climate averages for any city, has become a go-to resource but has been criticized for its smoothing algorithms. The new site could follow a similar trajectory—beloved by the public, scrutinized by scientists.
Looking ahead, the creator plans to add real-time alerts and a “climate shift” feature that compares current extremes to pre-industrial baselines. If successful, the tool could become a benchmark for how we communicate weather extremes in a warming world. But first, it needs to survive the roast.
