Facebook wants to learn more about where people around the world live and how they prefer to connect to the Internet. While there are hundreds of algorithms out there which provide detailed population information, none of these are enough to fulfill the social giant’s requirements. So, the company is doing the most logical thing. It is designing its own population maps.
Facebook’s Connectivity Lab took up this project along with the company’s data science division, infrastructure unit, and machine learning and artificial intelligence groups a few months ago. The company covered 21.6 million across 20 countries, which accounted for 14.6 billion images (over 350 TB of data) and analysed this graphical data to find evidence of human settlements.
The company used regular visual processing techniques including the image-recognition engine Facebook uses to identify people’s faces in photos and used deep learning to store data about human built structures.
The main aim of this project was to determine what Internet connection technique (mobile data or WiFi hotspots) is for bringing people online that would significantly increase the website’s traffic.
Our goal is to figure out how we can develop technologies and understanding how we can connect every person on the planet,
Yael Maguire, the director and head of Facebook’s Connectivity Lab said.
If we want to think about the best technology to connect people — especially outside of the cities, this is a very misleading piece of information.
The technology uses machine learning at its core, so it is apparent that it will get better with time and use. Facebook says that it has dedicated a few thousand machines from its data centers to run this service. The structures the algorithm recognized are laid out to determine where and under what conditions people live. The team then used census data to redistribute population evenly across each region assuming the best case scenario.
This results in the most detailed human population map you could get anywhere.
The data Facebook that Facebook gathers over the next few months using this technique will go live for global users sometime later this year. The company has already partnered up with Columbia University to learn the best approach for a global roll out of this service.
We believe this data has many more impactful applications, such as socio-economic research and risk assessment for natural disasters,
Facebook said in a statement.