Although opening data is important the format of this data is key to make it useful.

I usually hear people saying things like “they have opened some cool data, they rock!”. But when you look into that data what you see is a bunch of unrelated datasets, even partial data with some internal references to who-knows-what. So, it is a nightmare making that data useful. As a consequence these kinds of data are less used than expected. We need to reduce the consumption effort as much as possible.

What do you think about this?