Artificial intelligence meets weather: Time for change

Metswift, BIM software, Artificial Intelligence,

Construction weather and natural climate insight will change. Don’t get left out in the cold – BIM software can hot up via AI

High winds, extreme temperatures, torrential downpours and natural catastrophes. Globally, our increasingly erratic climate radically impacts construction planning and project management. MetSwift asks, how can we utilise newly emerged, rapidly-advancing Artificial Intelligence to take the pain out of weather prediction?

Here at MetSwift our data scientists, meteorologists, coders and business specialists have the inside take.

Increasingly extreme and erratic weather is costing construction

Extreme weather is a significant contributor to loses in construction. Costs associated with weather are as high as 21%. In March 2018, the “Beast from the East” cost the construction industry nearly £2bn – in three days.

What can be done to reduce these losses? We ask how existing and changing weather modelling can facilitate construction planning – both at short notice and far ahead of time.

The existing situation: BIM software and forecasting

Weather data currently available in BIM is simply historical weather data drawn from local weather stations, displayed graphically. However, there is no forecasting element actively integrated into 4D or 5D BIM, meaning this data is of limited value in the design and planning phases.

Long-range forecasting requires the skill and time of meteorologists, and sometimes climatologists. These weather professionals spend days researching a specific location then providing a written report of potential weather risk. This report – probably in the form of a PDF – lands on the desk, or in the inbox, of a build designer or project manager who makes an estimation on the time and cost impact in their 4D and 5D BIM planning.

With today’s advances in Artificial Intelligence, surely there are ways for construction to access weather insight that is faster, more accurate, and responsive to your points of concern; making processes more fluid and lessening loss.

It’ll let you down in the long run…

Weather services use some of the biggest computers in the world to do insanely complicated atmospheric physics to generate the weather forecasts we see after the evening news.

This form of numerical (or physics) modelling is geared toward giving incredible detail at short to medium range, but this method can’t be extrapolated in the long term.

Beyond two weeks, the accuracy is marginal and by three weeks, it shares very little in common with reality.

Instant, accurate, predictive, parametric. A new meteorological intelligence…

Our data scientists, meteorologists, coders and business specialists have developed AI algorithms that forecast far beyond the capabilities of existing numerical models.

We collected weather data from 92,456 accredited weather stations around the globe. That’s 98.8% coverage – we haven’t covered Yemen as it has no weather stations!

Our expert meteorologists processed and cleansed 13.68bn days of data, identifying anomalous and incorrect data pieces – based on their specialist knowledge of weather patterns and systems. This has actively, intricately trained our AI to identify specific weather events, their lifecycles and how they interact with each other; and this AI continues to learn from its own learning.

Through this machine learning and sizable computing power, our model then weights these events and scenarios to predict their effect on the weather, up to five years in advance. Via continual machine learning, it is continually testing, modifying inputs and improving its results to deliver accurate probability results for your thresholds for a specific weather event: at your selected location and date.

By replacing numerical modelling with statistically driven machine learning, we can provide remarkably accurate statistical probability percentages for your operational weather limits – and most notably extreme weather events – uniquely far in advance.

At these timescales, with this innovating intelligence – and at 98.8% of global coverage – our software finally provides forecasts useful in all aspects of the design and planning stages of construction projects, managed in the BIM software.

Making BIM better: Configuring weather, timely predictions

These predictions can be easily incorporated into BIM instantly. Once uploaded, various exact ‘what if’ scenarios are posed, or an intelligent form of. From these, real-time time cost projections are determined.

What if the crane will arrive on site 12 days later than initially planned? The revised timeline will automatically access the appropriate wind at height information, making responsive changes regarding time cost analysis.

Alternately, is it worth the extra cost in increased workforce and overtime to enclose structure sooner? This depends on the probability of extreme weather inhibiting work done exposed to elements.

Holistic and detailed, essential weather-based project-managerial issues: Now parametric probability modelling means decisions might be made with more confidence, and with a fascinating, flexible, precise, far higher degree of accuracy.

Meet our meteorologists at @MetSwift_Social … along with our business experts & data scientists.

From these advances…

BIM has seen great progress in recent years, but predictive weather elements have been left behind.

With the progress MetSwift has made in meteorological prediction, a game-changing paradigm shift in 4D and 5D BIM will be the incorporation of weather prediction:

Real-time changes in time and cost analysis that incorporates weather prediction is around the corner.

We have completed the intelligence…you assist the integration.



Tel: +44 (0)207 846 3277

+44 (0)755 778 2317


Please note: this is a commercial profile.


Please enter your comment!
Please enter your name here