Hubbry Logo
Weather Underground (weather service)Weather Underground (weather service)Main
Open search
Weather Underground (weather service)
Community hub
Weather Underground (weather service)
logo
7 pages, 0 posts
0 subscribers
Be the first to start a discussion here.
Be the first to start a discussion here.
Weather Underground (weather service)
Weather Underground (weather service)
Weather Underground (weather service)
View on Wikipedia
from Wikipedia

Weather Underground is a commercial weather service providing real-time weather information over the Internet. It provides weather reports for most major cities around the world on its Web site, as well as local weather reports for newspapers and third-party sites.

Key Information

Its information comes from the National Weather Service (NWS), and over 250,000 personal weather stations (PWS). The site is available in many languages, and customers can access an ad-free version of the site with additional features for an annual fee.

In February 2024, Weather Underground and its parent company, The Weather Company, became controlled by Francisco Partners.[1]

History

[edit]

The company is based in San Francisco, California and was founded in 1995 as an offshoot of the University of Michigan internet weather database. The name is a reference to the 1960s radical left-wing militant organization the Weather Underground, which also originated at the University of Michigan.[2][3] The group took its name from Bob Dylan's lyric "You don't need a weatherman to know which way the wind blows", from his 1965 song "Subterranean Homesick Blues".[4] The name, formerly UM-Weather, was changed to Weather Underground in 1991 due to feedback from the National Science Foundation[5] in response to Perry Samson's proposal for funding.[3] Weather Underground has since adopted the nickname Wunderground in addition to Weather Underground.

Jeff Masters, a doctoral candidate in meteorology at the University of Michigan working under the direction of Professor Perry Samson, wrote a menu-based Telnet interface in 1991 that displayed real-time weather information around the world. In 1993, they recruited Alan Steremberg and initiated a project to bring Internet weather into K–12 classrooms. Weather Underground president Alan Steremberg wrote "Blue Skies" for the project, a graphical Mac Gopher client, which won several awards. When the Mosaic Web browser appeared, this provided a natural transition from "Blue Skies" to the Web.

The original logo, used from 1997 through 2014

In 1995 Weather Underground Inc. became a commercial entity separate from the university.[6] It has grown to provide weather for print sources, in addition to its online presence. In 2005, Weather Underground became the weather provider for the Associated Press; Weather Underground also provides weather reports for some newspapers, including the San Francisco Chronicle. Alan Steremberg also worked on the early development of the Google search engine with Larry Page and Sergey Brin.[7]

In October 2008, Jeff Masters reported that the site was No. 2 for Internet weather information in 2008.[8]

In February 2010, Weather Underground launched FullScreenWeather.com (now a redirect to weather.com), a full screen weather Web tool with integrated mapping and mobile device use in mind.

On July 2, 2012, The Weather Channel announced that it would acquire Weather Underground, which would become operated as part of The Weather Channel Companies, LLC, which was later renamed "The Weather Company". The Weather Underground web site continued to operate as a separate entity from The Weather Channel primary site, weather.com, with its existing staff retained. Third-party web analytics providers Alexa and SimilarWeb rated the site as the 117th and 98th most-visited site in the United States, respectively, as of July 2015.[9][10] SimilarWeb rates the site as the second most visited weather website globally, attracting more than 47 million visitors per month.[10][11] The Weather Company also used the site's San Francisco headquarters as a regional office.[12][13]

The site's popularity also helped launch a television show hosted by meteorologist Mike Bettes, which originally aired on The Weather Channel from 1 p.m. to 5 p.m. ET. The show was renamed Weather Unfiltered on May 13, 2024 and moved to 6 p.m. to 10 p.m. as of October 29, 2024 as part of a major schedule adjustment for the network.[14]

On October 28, 2015, Jeff Masters noted that IBM had officially announced an agreement to acquire The Weather Company's mobile and cloud-based Web properties, including Weather Underground, WSI, weather.com, and also the Weather Company brand. The Weather Channel television service remained a separate entity, later sold to Entertainment Studios in 2018.[15] The deal was finalized on January 29, 2016.[16]

On October 3, 2019, Jeff Masters announced that he would be leaving Weather Underground.[17]

On August 22, 2023, IBM agreed to sell The Weather Company to private equity firm Francisco Partners for an undisclosed sum.[18] After the acquisition of the company was completed in February 2024, Weather Underground also became controlled by the American private equity firm.

Blogs

[edit]

Web logs (blogs) were one of the main features in Weather Underground, allowing users of the site to create blogs about weather, everyday life and anything else. Jeff Masters started the first blog on April 14, 2005,[19] and he posts blog entries nearly every day. From 2007 through early 2017 Richard B. Rood wrote blogs on climate change and societal response, with new entries on a weekly basis.

On October 14, 2016, the Wunderblog announced that it would be changing their name to Category 6, a name suggested by Jeff Masters. They decided on the name, because it "alludes to our deep fascination with all types of weather and climate extremes, including the many important facets of our changing climate", and "will provide all the insight and expert analysis needed to put the extreme events of our evolving 21st-century climate into context."[20]

On April 3, 2017 Weather Underground ended all Member blogs, WUMail, SMS alerts, NOAA Weather Radio rebroadcast and Aviation.[21]

Services

[edit]

Weather Underground also uses observations from members with automated personal weather stations (PWS).[22] Weather Underground uses observations from over 250,000 personal weather stations worldwide.[23]

The Weather Underground's WunderMap overlays weather data from personal weather stations and official National Weather Service stations on a Mapbox Map base and provides many interactive and dynamically updated weather and environmental layers.[24] On November 15, 2017, users were notified by email that their worldwide, user-provided weather cameras would cease to be available on December 15, 2017. However, on December 11, 2017 users received another email from Weather Underground announcing that they were reversing their position and would not be discontinuing the service based on significant user feedback.[25]

The service previously distributed Internet radio feeds of NOAA Weather Radio stations from across the country, as provided by users, and had a Weather Underground Braille Page.

The Associated Press uses Weather Underground to provide national weather summaries.[26]

Weather Underground has several Google Chrome extensions[27] and applications for iPhone, iPad and Android[28] including FullScreenWeather.com, a redirect to a full screen weather viewer tied into OpenStreetMap. There was an app developed for Roku devices, which has been deleted.[29]

In February 2015, Weather Underground released an iOS app called Storm.[30] This app is universal, and can be used on both iPhone and iPad. Other apps by Weather Underground include WunderStation[31] for iPad and WunderMap[32] for iOS and Android. In 2017, Weather Underground removed support for Storm, in favor of the "Storm Radar" app released by The Weather Channel Interactive in June 2017.[33]

On December 31, 2018, Weather Underground ceased offering its popular application programming interface (API) for weather data, further reducing the breadth of its services.[34]

On September 10, 2019, Weather Underground announced the discontinuation of its Email Forecast Program as of October 1, 2019, continuing the reduction in services noted above.[35]

See also

[edit]

References

[edit]
[edit]
Revisions and contributorsEdit on WikipediaRead on Wikipedia

Weather Underground (weather service)

View on Grokipedia
from Grokipedia

Weather Underground is an American online weather service founded in 1995 by Jeff Masters and colleagues at the , offering forecasts, historical weather archives, interactive radar maps, and alerts derived from data supplemented by observations from a global network of personal weather stations. Initially created as an extension of applications, the service pioneered internet-based dissemination, challenging traditional models by emphasizing user-accessible, data-rich interfaces and community-sourced inputs that enable granular, location-specific predictions. Its integration of stations has expanded coverage to remote and urban microclimates, fostering a collaborative model that enhances forecast accuracy through real-time, ground-level data aggregation. Acquired by The Weather Company in 2012, Weather Underground became part of a portfolio later purchased by IBM in 2016 for integration with cloud and analytics platforms, before The Weather Company's assets were sold to Francisco Partners in 2023, maintaining its operational focus on digital weather tools amid evolving corporate ownership. Notable for features like customizable alerts and historical query tools, it has garnered recognition for reliability in long-range and tropical forecasting, though it operates without significant public controversies tied to its service provision.

History

Founding and Early Years (1991–2000)

Weather Underground originated as a university-based weather information service at the . In May 1991, Jeff Masters, a alumnus with a B.S. (1982) and M.S. (1983) from the institution, launched UM-Weather using a Sun 4/110 workstation housed in the Space Research Building. This text-based system provided initial access to weather data via early internet protocols, attracting approximately 100 weekly users. The service was developed under the guidance of faculty advisor Perry Samson, an professor, with the aim of democratizing access to public-domain meteorological data from sources like the . Following feedback from the , which provided funding for educational enhancements, the service was renamed in 1991; Samson selected the name, drawing inspiration from a 1960s student activist group to evoke grassroots accessibility. Usage surged after made landfall on August 19, 1991, elevating weekly visitors to 25,000 as users sought real-time storm updates. Contributions from students Alan Steremberg (B.S.E. 1994) and Chris Schwerzler (B.S.E. 1996), along with staff member Jeff Ferguson, expanded its technical capabilities during this period. By 1993, the service introduced a client featuring interactive weather maps tailored for K-12 education, securing an Apple "Cool Tool" award for innovation in visual data presentation. In 1995, it transitioned to the as wunderground.com, initially offering daily forecasts and hourly conditions for 550 U.S. cities, marking a shift from text-only to graphical interfaces. The platform emphasized data aggregation from government sources, prioritizing accuracy over commercial forecasts, though it remained a small operation focused on educational and hobbyist users through the decade. By the late , incorporation as The Weather Underground, Inc., enabled a pivot to ad-supported free access, broadening its appeal amid growing adoption.

Expansion and Technological Advancements (2001–2011)

During the early 2000s, Weather Underground significantly expanded its personal weather station (PWS) network, launching the feature in 2001 to enable users worldwide to submit localized , thereby enhancing hyper-local granularity beyond traditional sources. This initiative marked a pivotal technological advancement, integrating community-sourced observations with official meteorological to provide denser coverage in underserved areas, with the network growing to 17,000 stations by 2011. The platform's user base surged amid rising internet adoption, achieving 10-20% annual growth even through the 2009-2010 recession, culminating in 17 million monthly visitors worldwide (13 million in the U.S.) by , ranking it 77th in U.S. web traffic. Operational expansion paralleled this, with employee headcount increasing to 38 by , including meteorologists, developers, and support staff distributed between and Ann Arbor. Revenue diversification supported scaling, deriving approximately 70% from advertising, 20% from custom data feeds to partners like and the , and 5% from newspaper syndication. Technologically, the period saw enhancements in user interactivity and data visualization, including the introduction of weather stickers for embeddable site widgets, user-uploaded photos exceeding 1.3 million by 2011, and blogging capabilities that drew thousands of comments per post, exemplified by chief meteorologist Jeff Masters' Wunderblog launched around 2005 for in-depth analysis. These features fostered a community-driven , while preparations for mobile applications underscored adaptation to emerging handheld devices by the decade's end. By 2011, plans for international expansion included localized , , and alerts to capitalize on global demand.

Acquisitions and Corporate Integration (2012–Present)

In July 2012, Companies, owned by a including , Blackstone Group, and , acquired Weather Underground to bolster its digital weather data capabilities and integrate features such as the WunderMap visualization tool. The acquisition preserved Weather Underground's independent brand and data-centric approach amid user backlash over potential commercialization, with assurances that core operations would remain data-focused rather than ad-driven. On October 28, 2015, IBM announced its intent to acquire The Weather Company's business-to-business, mobile, and cloud-based digital properties, explicitly including alongside weather.com and WSI, with the deal closing on January 29, 2016. This move integrated Weather Underground's forecasting data and personal weather station network into 's Watson platform for applications, enabling advanced analytics for enterprise clients in sectors like , , and by combining weather models with . Post-acquisition, Weather Underground's APIs and datasets supported IBM's expansion of weather-informed AI services, though consumer-facing features saw gradual enhancements in mobile apps and precision alerts without fundamental overhauls to its user-driven model. In August 2023, agreed to divest assets, including , to private equity firm in a transaction finalized by 2024, shifting ownership away from IBM's direct control after eight years of technological synergy. This sale decoupled from IBM's Watson ecosystem, allowing renewed emphasis on standalone digital consumer tools and monetization under ' portfolio strategy, which prioritizes scalable data platforms over broad enterprise integration. As of 2025, corporate integration has stabilized around hybrid data aggregation, with leveraging its legacy personal station network alongside third-party feeds, independent of prior IBM-specific AI dependencies.

Ownership and Business Model

Key Acquisitions and Ownership Changes

In July 2012, Weather Underground was acquired by The Weather Channel Companies (TWCC), a owned by NBCUniversal, Blackstone Group, and , marking its transition from an independent entity to integration within a larger weather media portfolio. This deal, announced on July 2, 2012, aimed to combine Weather Underground's personal weather station network and data with TWCC's broader forecasting capabilities and audience reach. On October 28, 2015, IBM announced its acquisition of 's digital and B2B assets, including , weather.com, and associated mobile and cloud properties, for a reported value exceeding $2 billion, though the television assets were retained separately. The transaction closed on January 29, 2016, integrating into 's ecosystem to leverage its data for Watson AI enhancements and enterprise analytics, while preserving operational independence under branding. In August 2023, IBM agreed to divest The Weather Company assets, including Weather Underground, to , a technology-focused , with the deal completing on February 1, 2024. This shift positioned , encompassing Weather Underground, as a standalone entity led by CEO Sheri Bachstein, emphasizing continued focus on data-driven weather services amid evolving enterprise demands. No subsequent major ownership changes have been reported as of October 2025.

Revenue Streams and Monetization Strategies

Weather Underground's primary revenue stream for its consumer-facing services derives from digital displayed on its website and mobile applications, where users access free weather forecasts, maps, and historical data. This ad-supported model leverages high traffic volumes, with targeted ads informed by location-specific weather data to enhance relevance for advertisers. As part of , these efforts integrate with broader weather-powered advertising solutions that enable brands to deliver geo-targeted campaigns, driving engagement through predictive weather triggers. A significant B2B monetization strategy involves paid access to the API, which supplies developers, enterprises, and third-party applications with real-time and historical weather data, including observations from its personal weather station network. Initially offering free API keys to foster ecosystem growth, the service discontinued this in May 2018 to prioritize enhanced user relationships and service improvements, transitioning to tiered paid plans that generate licensing fees based on usage volume. This shift supports for industries such as , , and , contributing to data-as-a-service (DaaS) revenues. Following its 2012 acquisition by and subsequent ownership changes—including IBM's 2016 purchase and ' 2024 acquisition—Weather Underground's strategies have aligned with enterprise-focused licensing and customized data products. These include syndicating proprietary datasets, such as crowdsourced personal weather station readings exceeding 250,000 units globally, to corporate clients for analytics and risk modeling, though specific contributions to overall revenues remain integrated within portfolio rather than itemized separately.

Data Sources and Forecasting Methodology

Personal Weather Station Network

The Personal Weather Station (PWS) network aggregates data from over 250,000 user-operated sensors worldwide, establishing Weather Underground's position as operator of the largest crowdsourced weather observation system. These stations, typically installed by hobbyists, homeowners, and small organizations, measure variables such as air temperature, relative humidity, and speed, barometric pressure, and rainfall accumulation, transmitting readings at intervals as frequent as every few minutes via internet-connected hardware. Compatible devices include models from manufacturers like Ambient Weather, Davis Instruments, and RainWise, which integrate directly through protocols enabling automated uploads. Established as a core feature since Weather Underground's inception in , the PWS network leverages voluntary contributions to fill gaps in official meteorological coverage, particularly in suburban, rural, and urban microclimates underserved by government stations. By 2014, the network encompassed more than 34,000 active stations, reflecting growth driven by declining costs of consumer-grade sensors and rising public interest in localized . Expansion continued into the , surpassing 250,000 stations amid broader adoption of IoT-enabled devices, with users accessing personalized dashboards for real-time visualization, graphing, and historical archiving through the WunderStation application. Weather Underground applies rigorous protocols to PWS submissions, including automated checks for implausible values (e.g., temperatures exceeding physiological limits or sudden jumps inconsistent with physics) and statistical validation against nearby official observations, discarding or flagging anomalous data to maintain . This processed input supplements proprietary models and feeds from entities like the , enabling granular forecasts that resolve variations over distances as short as hundreds of meters—such as urban heat islands or sheltered valleys—where sparse professional networks fall short. Empirical assessments indicate the network's density correlates with improved short-term predictive precision in non-standard locales, though individual station accuracy varies with factors like , exposure to direct , or siting errors (e.g., rooftop placement inflating readings).
Key Network MetricsValueSource
Total Stations>250,000
Data Variables TrackedTemperature, , , ,
Growth Example (2014)>34,000 stations
Update FrequencyUp to every few minutes

Integration of Proprietary and Third-Party Data

Weather Underground's forecasting system blends from its extensive network of over 250,000 personal weather stations (PWS), which provide , community-sourced observations of temperature, precipitation, wind, and other variables. This PWS , collected via user-hosted sensors and quality-controlled for reliability, forms a core asset that enables granular, neighborhood-scale insights not readily available from traditional sources. Third-party data integration supplements PWS inputs with broader observations, including approximately 180,000 U.S. stations (such as automated surface observing systems), 29,000 international stations, and 11,000 volunteers from the National Oceanic and Atmospheric Administration's (NOAA) Cooperative Observer Program (COOP), which contribute standardized daily measurements. Additional third-party feeds encompass networks for mapping and specialized partnerships, such as the U.S. Department of Agriculture's SNOTEL system with over 600 high-elevation sites in the western U.S. for monitoring. These external datasets, often from government agencies like NOAA and the , provide validated baseline conditions and regional context to mitigate gaps in PWS coverage. The integration process occurs through Weather Underground's proprietary algorithms, which assimilate PWS hyperlocal readings with third-party model outputs—such as from sources including NOAA's —to generate refined short-term forecasts emphasizing local variability. This hybrid approach prioritizes empirical station density for current conditions and historical records, while cross-validating against official observations to enhance overall precision, particularly in urban or topographically complex areas where PWS density is high. Post-2015 acquisition by (an IBM subsidiary), the system has incorporated advanced computational resources, but retains emphasis on PWS-driven personalization over purely model-based predictions.

Algorithmic and AI-Driven Forecasting

Weather Underground's proprietary forecasting system employs algorithms that assimilate from over 250,000 personal weather stations to generate predictions, blending local observations with broader numerical weather models for enhanced precision at the neighborhood scale. This approach prioritizes data-driven adjustments to account for microclimatic variations, supported by oversight to validate model outputs. Post-2015 acquisition by , an subsidiary, Weather Underground's forecasts integrate via the Deep Thunder system, which leverages historical data and personal station inputs to train predictive models for short-term, high-resolution local weather impacts. Deep Thunder utilizes to produce forecasts at 3-kilometer resolutions, employing AI for and scenario simulation beyond traditional physics-based simulations. Further advancements include situation-dependent for model blending, drawing on Weather Underground's unique dataset access to optimize forecasts dynamically. The Global High-Resolution Atmospheric Forecasting () model, operational since updates in 2025, incorporates AI techniques to refine atmospheric simulations with frequent updates and higher resolutions, underpinning Weather Underground's outputs. In 2024, collaborations with expanded AI capabilities, enabling faster processing of observational data for global-scale predictions integrated into Weather Underground's platform. These AI-driven elements improve computational efficiency and accuracy by identifying subtle patterns in large datasets that deterministic models might overlook, though details limit full transparency on implementation.

Core Services and Features

Website and User Interface

The Weather Underground website, hosted at wunderground.com, functions as the core digital platform for accessing weather data, forecasts, and interactive tools powered by over 250,000 personal weather stations. Its emphasizes data granularity and customization, delivering current conditions, extended forecasts, and historical trends tailored to specific locations via station proximity mapping. A significant redesign launched on April 15, 2014, introduced a responsive layout optimized for seamless functionality across desktops, tablets, and mobile devices, with content organization refined through direct user feedback to enhance and feature prioritization. The interface integrates interactive elements such as high-resolution overlays, , and severe weather alerts on customizable maps, supporting light and dark modes alongside varied map styles for user preference. For personal weather station owners, the Wunderstation dashboard provides specialized UI components, including tools to view, analyze, and compare local through configurable charts, graphs, and sharing options, promoting detailed meteorological visualization. Additional interface features encompass real-time monitoring panels, weather camera feeds, and layered maps, enabling users to overlay multiple types for comprehensive environmental assessment. This design prioritizes empirical presentation over aesthetic simplicity, reflecting the service's foundation in crowdsourced observations for precise, location-specific insights.

Mobile Applications and API Access

The Weather Underground mobile applications are available for and Android devices, delivering hyper-local forecasts powered by data from over 250,000 personal weather stations alongside proprietary modeling. Key features include interactive maps, high-resolution , alerts, customizable overlays such as heat maps and personal station data, and support for light/dark modes with varied map styles. The version, "Weather Underground: Local Map," holds a 3.9 out of 5 rating from 31,090 user reviews on the , emphasizing its precision for localized conditions like air quality and UV index. On Android, the app "Weather data & : Weather Underground" achieves a 4.7 rating from 602,834 reviews, with over 10 million downloads reported, highlighting its utility for tracking and forecast accuracy. API access for developers enables programmatic retrieval of Weather Underground data, including current conditions, forecasts, details, and historical records via authenticated requests requiring an . Originally offering a free tier popular among non-commercial users, the service transitioned to paid plans in December 2018, with entry-level commercial access starting at $850 monthly following the end of free public endpoints. A restricted low-volume option for personal, non-commercial applications was subsequently provided to mitigate disruption for hobbyist integrations. Post-acquisition by in 2015, API functionalities have increasingly aligned with enterprise-grade offerings, prioritizing high-volume, paid developer use over broad free access.

Specialized Tools and Content (e.g., Historical Data, Blogs)

Weather Underground provides access to historical weather through its Weather History & Data Archive, enabling users to retrieve past observations by specifying a location via city, , or airport code, along with a specific date. This tool compiles including , , , , and from official stations and the site's extensive personal weather station (PWS) network, with records dating back decades in many areas. Users can view daily summaries, hourly breakdowns, or graphical charts, supporting applications like , , and event planning. The platform's PWS integration enhances historical content by archiving user-submitted data from thousands of stations, allowing station owners to access and export their long-term records for analysis. This crowdsourced historical repository supplements traditional meteorological sources, offering hyper-local granularity not always available from national agencies. In terms of content, Weather Underground formerly emphasized user-generated WunderBlogs, where individuals could post entries on weather events, personal observations, and related topics, fostering community-driven discussions. Professional blogs, such as Category 6 by meteorologists Dr. Jeff Masters and Bob Henson, provided in-depth analysis of hurricanes, climate trends, and extreme weather from 2013 until its discontinuation in 2020, after which content was archived or relocated. The site maintains a news section for curated weather updates and articles, though it has shifted toward more standardized reporting post-2015 acquisition by The Weather Company. These features historically differentiated Weather Underground by blending data tools with interpretive content, though user blogging has diminished in prominence.

Accuracy, Reception, and Criticisms

Empirical Accuracy Evaluations

Independent evaluations of Weather Underground's forecast accuracy, primarily conducted by ForecastWatch using datasets encompassing millions of forecasts, have shown competitive performance relative to major providers like , , and government sources such as NOAA. ForecastWatch's analyses, which measure metrics including error (RMSE) for temperature forecasts and (POP) accuracy, indicate that Weather Underground ranked among the top performers in several categories from 2010 onward. For instance, in evaluations of one- to five-day-out global low temperature forecasts from January to June 2016, Weather Underground tied for first place with and . Similarly, for U.S. high temperature forecasts in the same period, it tied for second behind . Following its 2012 acquisition by (now under IBM's ), Weather Underground's forecasting methodology integrated more closely with TWC's systems, resulting in nearly identical accuracy scores from late 2013. A ForecastWatch three-region accuracy overview spanning 2010-2017 found Weather Underground to be the most accurate provider in 12 of the evaluated periods across high and low forecasts, outperforming competitors in regional subsets. It also demonstrated the largest improvement in overall accuracy, gaining 4.60 percentage points from 2013 to 2016, attributed to enhanced and model refinements. These gains were evident in short-term high forecasts, where Weather Underground achieved lower RMSE errors compared to earlier years. Precipitation forecasting evaluations reveal more variability. In long-term analyses of short-term POP forecasts, Weather Underground's performance aligned with industry leaders but lagged in wind speed predictions, where it ranked last alongside The Weather Channel, exhibiting 47.2% lower accuracy than top performers in a 2020 study cited by . ForecastWatch's POP studies from 2014-2015 highlighted Weather Underground's reliability for short-term events, though overall rankings placed it behind specialized providers in extended-range probability. These metrics underscore strengths in temperature prediction—core to daily forecasts—while noting challenges in probabilistic elements like and rain, common across commercial services due to inherent atmospheric chaos. Comparisons with NOAA, the primary public benchmark, show Weather Underground often surpassing government models in commercial evaluations, though NOAA's raw data feeds many private forecasts, including post-acquisition enhancements. ForecastWatch data from 2006-2016 for one-day-out temperatures confirmed Weather Underground's edge in localized accuracy, benefiting from its personal weather station network for ground-truth validation. However, self-reported claims by competitors like and , which tout superior AI-driven edges in recent years, warrant scrutiny for promotional bias, as independent verifications like ForecastWatch's predate widespread AI adoption in models. Empirical limitations persist: evaluations typically cover contiguous U.S. regions, with sparser global data, and accuracy degrades predictably beyond five days due to model , a universal constraint rooted in numerical weather prediction's sensitivity to initial conditions.

User and Expert Reviews

Users have expressed mixed sentiments toward Weather Underground's services, with app store ratings generally higher than aggregate review site scores. On the Apple , the app holds a 3.9 out of 5 rating from over 31,000 reviews as of recent data, with praise for its integration, aviation layers, and hyperlocal data from personal weather stations, though some criticize the interface as overly cluttered with extraneous features. In contrast, Google Play users rate it 4.7 out of 5 from more than 602,000 reviews, appreciating its focus on essential weather metrics without unnecessary bloat and the value of ad-removal subscriptions, positioning it as a straightforward tool for core forecasting needs. However, dedicated review platforms reflect greater dissatisfaction; scores it 1.3 out of 5 from 86 reviews, where former enthusiasts lament post-acquisition interface overhauls that degraded usability and forecast reliability, describing it as transitioning from a preferred service to one of the least effective options. Sitejabber similarly reports a 1.4 out of 5 rating from 667 reviews, with widespread complaints about persistent bugs, intrusive ads, and diminished hyperlocal precision following the 2015 acquisition. Expert evaluations highlight Weather Underground's strengths in specialized data aggregation while noting usability drawbacks. PCMag reviewers commend its smart forecasting algorithms, detailed storm tracking, air quality metrics, and seamless links to extensive website content, recommending it for users seeking granular, station-based insights over generalized predictions. Comparative analyses, such as those from Weatherstack, position it as a solid contender against apps like AccuWeather and The Weather Channel, valuing its proprietary personal weather station network for microclimate accuracy but critiquing slower load times and less intuitive mobile experiences compared to rivals. Aviation and meteorology enthusiasts, per forums like Pilots of America, decry post-2018 redesigns that removed key features such as full Nexrad radar access, rendering it less viable for professional operational use despite its historical edge in crowd-sourced data. Overall, experts attribute its enduring appeal to the vast personal station dataset—over 250,000 units globally—but fault corporate integrations under IBM for prioritizing monetization over the lean, data-centric model that defined its pre-acquisition reputation.

Criticisms of Post-Acquisition Changes

Following IBM's acquisition of The Weather Company's digital assets, including , in October 2015, users criticized the service for degraded performance and usability. Reports highlighted slow loading times, frequent app crashes, and website timeouts, with one user noting the need to "load it, close and reopen the app about 95% of the time" to access data. UI redesigns were faulted for prioritizing aesthetics over functionality, replacing simple, data-dense graphs with slower interfaces that users described as "useless." These issues contributed to app ratings dropping to around 2.5 stars on platforms like by 2020. Increased commercialization drew further backlash, as the platform incorporated more advertisements and headlines, shifting from its origins as a , user-driven data aggregator. Users on forums reported the app becoming "full of horrible links," diluting the core value of personal integrations and historical data access. In October 2021, the removal of the service eliminated a key feature for real-time visual verification, prompting complaints that essential tools were being stripped to cut costs. Feature discontinuations amplified perceptions of neglect. The public , vital for third-party developers and , ended on December 31, 2018, forcing migrations to less accessible alternatives. Layoffs in May 2020 at , affecting dozens including content creators, resulted in the June closure of the Category 6 blog, a respected source for advanced meteorological analysis by experts like Jeff Masters and Bob Henson. Some users attributed these cuts to IBM's focus on enterprise data sales over consumer-facing improvements, leading to a broader sense of service decline despite corporate claims of AI-enhanced accuracy. Individual reports of forecast errors, such as temperature discrepancies of 10°F or underestimated wind speeds, fueled skepticism, though these remained anecdotal amid vendor-verified global accuracy gains.

Impact and Legacy

Contributions to Weather Forecasting

Weather Underground advanced through the development of a system that integrates observations from its extensive personal weather stations (PWS) network with traditional data sources, including cooperative observer program (COOP) stations, airports, and weather balloons, to produce neighborhood-scale predictions. Launched in as an extension of a weather database, the service initially offered hourly forecasts and severe weather information for over 550 cities, making detailed outputs accessible via the internet ahead of many competitors. The PWS network, which expanded to over 250,000 stations worldwide by the , captures real-time variations—such as urban heat islands or localized —that official sparse networks often overlook, enabling forecasts with finer down to the block level. Strict protocols filter these crowdsourced inputs to mitigate errors from uncalibrated sensors or siting issues, ensuring reliable assimilation into models. Empirical analyses confirm the value of this approach: quality-controlled Weather Underground PWS data, when applied in urban reanalysis, yields measurable accuracy improvements in air temperature estimates through sequential filtering techniques that remove outliers and biases. Similarly, incorporating PWS observations enhances mapping by densifying gauge coverage, reducing interpolation errors in nowcasting and short-term forecasts. These enhancements stem from the causal role of increased observational in constraining model uncertainties, particularly in heterogeneous terrains. Additionally, Weather Underground's archival of historical weather from PWS and other sources facilitates verification of forecast models and supports probabilistic improvements over time, contributing to iterative refinements in ensemble predictions. By prioritizing -driven hyperlocalization over generalized regional outputs, the service has demonstrated how citizen-contributed observations can operationally bridge gaps in professional networks, influencing subsequent adoptions of similar hybrid strategies in commercial and research forecasting.

Influence on Competitors and Industry Standards

Weather Underground's pioneering aggregation of data from personal weather stations (PWS) established a model for crowdsourced that competitors emulated to achieve greater resolution in predictions. By the early , the service had integrated observations from over 40,000 user-operated stations worldwide, allowing for microclimate-specific reports that traditional grid-based models could not match, thereby demonstrating the feasibility and accuracy benefits of community-sourced inputs. This innovation compelled rivals, including and early iterations of The Weather Channel's digital platforms, to expand their own localized data collection efforts, fostering an industry-wide shift toward incorporating amateur and IoT sensor networks for enhanced granularity in urban and rural areas. The service's emphasis on accessible, data-rich APIs and historical archives further influenced standards for weather app development, where developers increasingly prioritized real-time PWS integration for competitive differentiation. Post-acquisition by Companies on July 2, 2012, Weather Underground's PWS infrastructure—now exceeding 250,000 stations—bolstered the acquirer's capabilities, underpinning advancements such as the 2016 Deep Thunder model, which fused observations with high-resolution simulations to predict convective events like thunderstorms at sub-kilometer scales. These contributions elevated baseline expectations for forecast detail and reliability, evident in the subsequent proliferation of similar features across enterprise tools and consumer apps, including AI-augmented outputs from providers like IBM's weather division.

References

  1. https://lsa.umich.edu/umbs/news-events/all-events.detail.html/30830-3794861.html
  2. https://www.wunderground.com/
  3. https://www.wunderground.com/pws/overview
  4. https://leadiq.com/c/weather-underground-acquired-by-the-weather-company-august-2012/5a1d98fa23000052008792c8/employee-directory
  5. https://news.engin.umich.edu/2017/02/wunderground-com-democratizing-weather/
  6. https://www.prnewswire.com/news-releases/ibm-closes-deal-to-acquire-the-weather-companys-product-and-technology-businesses-300212259.html
  7. https://www.weathercompany.com/news/francisco-partners-to-acquire-the-weather-company-assets-from-ibm/
  8. https://yaleclimateconnections.org/2010/09/wundergroundcom-cofounder-jeff-masters/
  9. https://michigandaily.com/uncategorized/university-alums-sell-weather-underground-weather-channel-after-17-years-independence/
  10. https://www.wunderground.com/about/our-company
  11. https://concentratemedia.com/jeffmastersweatherunderground00142/
  12. https://www.blackstone.com/news/press/the-weather-channel-companies-acquire-weather-underground/
  13. https://www.nytimes.com/2012/07/04/us/as-weather-channel-buys-weather-underground-fans-fear-change.html
  14. https://uk.newsroom.ibm.com/2015-Oct-28-IBM-Plans-to-Acquire-The-Weather-Companys-Product-and-Technology-Businesses-Extends-Power-of-Watson-to-the-Internet-of-Things
  15. https://www.engadget.com/2015-10-28-ibm-weather-company-acquisition.html
  16. https://newsroom.ibm.com/2023-08-22-Francisco-Partners-to-Acquire-The-Weather-Company-Assets-from-IBM
  17. https://www.cnbc.com/2023/08/22/ibm-sells-the-weather-channel-and-the-rest-of-its-weather-business.html
  18. https://maximosecrets.com/2017/03/21/why-did-ibm-acquire-the-weather-company/
  19. https://www.wunderground.com/blog/JeffMasters/wundergroundcom-sold-to-the-weather-channel-companies.html
  20. https://weather.com/news/news/twc-weather-underground-20120702
  21. https://www.franciscopartners.com/media/francisco-partners-completes-acquisition-of-the-weather-company
  22. https://www.weathercompany.com/advertising/
  23. https://vizologi.com/business-strategy-canvas/the-weather-company-business-model-canvas/
  24. https://forum.pvoutput.org/t/weather-underground-api-gone/576
  25. https://www.visualcrossing.com/resources/blog/replacing-the-weather-underground-api/
  26. https://www.wunderground.com/about/data
  27. https://www.wunderground.com/pws/buying-guide
  28. https://www.usatoday.com/story/tech/personal/2014/04/15/weather-underground-gets-new-look/7715699/
  29. https://www.wunderground.com/wunderstation/
  30. https://community.netweather.tv/topic/98966-weather-station-accuracy-vs-price/
  31. https://play.google.com/store/apps/details?id=com.wunderground.android.weather&hl=en_US
  32. https://www.instrumentchoice.com.au/blogs/news/what-is-weather-underground
  33. https://d3.harvard.edu/platform-rctom/submission/will-ibms-deep-thunder-finally-be-the-key-to-accurate-weather-forecasting/
  34. https://uk.newsroom.ibm.com/2016-Jun-15-The-Weather-Company-Launches-Deep-Thunder-the-Worlds-Most-Advanced-Hyper-Local-Weather-Forecasting-Model-for-Businesses
  35. https://www.informationweek.com/machine-learning-ai/ibm-s-deep-thunder-shows-how-weather-is-big-business
  36. https://www.weathercompany.com/global-high-resolution-atmospheric-forecasting/
  37. https://www.weathercompany.com/news/twco-adopts-jedi-to-improve-everyday-forecasts/
  38. https://www.weathercompany.com/news/the-weather-company-expands-collaboration-with-nvidia-to-advance-ai-based-weather-forecasting-and-visualization-capabilities/
  39. https://ci.uky.edu/irj/rural-blog/weather-reports-future-could-be-faster-and-more-accurate-help-ai-forecasting-tools
  40. https://www.weathercompany.com/blog/ai-in-weather-forecasting-prediction-and-communication/
  41. https://www.wunderground.com/download
  42. https://columbiaweather.com/resources/setting-up-weatherunderground/
  43. https://www.wunderground.com/blog/JeffMasters/wunderground-launches-major-site-redesign-and-new-logo.html
  44. https://www.wunderground.com/maps
  45. https://apps.apple.com/us/app/weather-underground-local-map/id486154808
  46. https://play.google.com/store/apps/details?id=com.wunderground.android.weather
  47. https://play.google.com/store/apps/details?id=com.wunderground.android.weather&hl=en_SG
  48. https://www.powercmd.com/api-wunderground/
  49. https://www.reddit.com/r/webdev/comments/8tjavu/now_that_the_free_wunderground_api_has_been/
  50. https://community.home-assistant.io/t/end-of-service-for-the-weather-underground-api/70072
  51. https://www.omi.me/blogs/api-guides/how-to-fetch-weather-forecast-using-weather-underground-api-in-javascript
  52. https://www.wunderground.com/history/
  53. https://www.wunderground.com/blog/
  54. https://www.wunderground.com/cat6
  55. https://www.wunderground.com/news
  56. https://forecastwatch.com/wp-content/uploads/Global_Temperatures_One_to_Five_Day_Out_Jan-June_2016.pdf
  57. https://forecastwatch.com/wp-content/uploads/USHighTempsJanJun2015.pdf
  58. https://forecastwatch.com/wp-content/uploads/Three_Region_Accuracy_Overview_2010-2017.pdf
  59. https://forecastwatch.com/wp-content/uploads/Three_Region_Overview_2010_201606.pdf
  60. https://forecastwatch.com/wp-content/uploads/HighTemps2006-2015.pdf
  61. https://www.prnewswire.com/news-releases/latest-independent-study-reconfirms-superior-accuracy-of-accuweather-forecasts-300981873.html
  62. https://forecastwatch.com/wp-content/uploads/Short_Term_POP_Accuracy_2014.pdf
  63. https://www.forecastwatch.com/wp-content/uploads/Short_Term_POP_Accuracy_2015_Ag.pdf
  64. https://forecastwatch.com/wp-content/uploads/HighTemps2006-2016.pdf
  65. https://www.weathercompany.com/news/twco-widens-lead-as-worlds-most-accurate-forecaster/
  66. https://www.accuweather.com/en/press/latest-study-of-120-million-forecasts-proves-accuweather-forecasts-are-most-accurate/662380
  67. https://www.trustpilot.com/review/www.wunderground.com
  68. https://www.sitejabber.com/reviews/wunderground.com
  69. https://www.pcmag.com/picks/best-weather-apps
  70. https://weatherstack.com/which-weather-app-is-the-most-accurate
  71. https://www.pilotsofamerica.com/community/threads/weather-underground-is-now-useless.107999/
  72. https://news.ycombinator.com/item?id=22740997
  73. https://www.reddit.com/r/weather/comments/15y8wkh/ibm_to_sell_their_weather_channel_division/
  74. https://www.reddit.com/r/weather/comments/qb554l/wunderground_has_gone_from_great_to_decent_to/
  75. https://community.openhab.org/t/weather-underground-end-of-service-announcement/51414
  76. https://www.niemanlab.org/reading/weather-company-is-shuttering-its-popular-weather-underground-category-6-blog-after-layoffs-at-parent-company-ibm/
  77. https://newsroom.ibm.com/2021-07-29-IBMs-The-Weather-Company-Continues-to-Be-the-Worlds-Most-Accurate-Forecaster-Overall
  78. https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2019GL083731
  79. https://hess.copernicus.org/articles/25/583/2021/
  80. https://www.reddit.com/r/explainlikeimfive/comments/1f4921z/eli5_what_is_weather_underground_and_why_does/
  81. https://www.prnewswire.com/news-releases/the-weather-company-launches-deep-thunder---the-worlds-most-advanced-hyper-local-weather-forecasting-model-for-businesses-300285177.html
Add your contribution
Related Hubs
User Avatar
No comments yet.