The article also seems to presume that there is some single source where the data from all these disparate systems is tied together, but no such thing exists afaik.
You can see that this is the problem in many places. One of those places which springs to mind is that, for example, Google argues that they are just like any user of the (infrastructure of the) internet. But whenever two individuals interact via anything Google, the communication goes mediated by Google. In a sense it is centralized, and there is a centralized map, even if physically spread between variously located data centers.
There are multiple billion-dollar companies whose entire business is tying together data from all these disparate systems, e.g. Palantir. There are leaked multiple government programs whose entire purpose is the same, e.g. XKeyscore. Or look at something more general purpose like Splunk, where it is trivial to filter 1TB of logs from different sources and map identifiers.