We accessed real time data on wildfires, record temperatures, air quality and CO2 levels from satellites and weather monitoring systems around the world and used it to create geolocated alerts on Earth’s twitter account. Once the matrix was generated, containing the potential natural disaster, illustrating image and tagging the people and organizations responsible to make a change about it, and thanks to machine learning, posts needed no human intervention.

All the natural disasters that we witness are more than that, it´s planet earth trying to communicate with us, messages in the form of hurricanes, glacial melt, highest or lowest temperatures ever, sandstorms, and all sorts of natural phenomena. But what is the message behind this? its planet earth telling us to stop throwing plastic to the oceans, to stop starting fires in the Amazon rainforest, or Australia's explanadle, the Patagonia or California, to reduce the air pollution that deteriorates the ozone layer, so we can care for the air we all breathe. All of this is what planet Earth is trying to make us aware of with its language. Even worse, that information is sometimes kept away from the general public or it's manipulated.

For all of these reasons it was about time to get the message straight, for us to take action, before it is too late.

We gave earth its own Twitter account @EarthIsSaying and synchronized it with satellites and environmental monitors around the world. Using Machine Learning, we converted data from floods, forest fires, air quality, temperature and CO2 levels, into tweets from the earth in real time. The alert determines not only the message, but also precise data on real time to illustrate the potential natural disasters and give it a context, but also tagged the relevant people that should take action. In that way earth itself without human intervention, made its health condition a trend among its inhabitants and leaders around the world. Reaching younger generations and not only creating a conversation on twitter, but urging to take action before it´s too late.

Using the world's most advanced environmental monitoring systems: IQAir, Ambee, Center for Polar Observation and Monitoring and AccueWeather among others, we needed to reach out to a non-environmentalist younger generation in a distinctive way, to encourage with Greenpeace their real concern for the planet.

To achieve it we created the first Twitter account for the earth that sends messages without human intervention. An account aimed to make entities, personalities and local and national governments, see everything the planet is going through at the exact time that environmental alerts for fires, floods, CO2 levels and air quality, are happening. All of this messages to help them take action and find solutions that build everyone’s care for our environment.

We collected data on floods, fires, CO2 levels, record temperatures and air quality taken from satellites and climate sensors around the world in real time, and we crossed it with the location of the alert through geolocation and the entities involved, to result in tweet-like messages tagging the incumbent authorities that could help solve the situation. We were able to give the exact number in which emissions were exceeded and the amount of times it had happened in recent years for example.

We combined:

Environmental Issue + Alert + Location + @ + Message = Tweet without human intervetion. And with the use of machine learning we just got better and better each tweet.

Results on Earth’s day:

• Verified by Twitter – (Non human intervention account).

• + 171,2 MM impressions

• Trending topic in an hour on launch date

• 570 early alerts generated on potential natural disasters

• 43 countries interacted:

Colombia, España, Chile, Argentina, Bangladesh, Italy, Afghanistan, Japan, Indonesia, Nepal, Vietnam, China, Bulgaria, South Korea, Peru, Malasia, UAE, Iceland, India, Mexico, Russia, USA, Kuwait, Israel, Egypt, Singapur, England, South Africa, Canada, Hungary, France, Iran, Uganda, Pakistan, Bahrein, Romania, Eslovaquia, Phillipines, Mongolia, Bosnia-Herzegovina, Macedonia, Ghana.

• 12 languages tweeted