All datasets have a very important part in machine learning and as software becomes more of a commodity and it becomes easier for people to make software. The only moat that people could have is high-quality data and the amount of data so the more data you have the more improved and intelligent software will probably be. I guess it's a very good initiative for the community to get access to data. I just wanted to ask: what kinds of data does it covers? Does it have mechanical or physical data as well?