It means that in order to give a topological space, you need to give two pieces of information:
1) First, you need to give an underlying set ##X##
2) Secondly, you need to give a set ##\mathcal{T}## which consist of subsets of ##X##. This set ##\mathcal{T}## is called the topology. It can be completely arbitrary, but it does need to satisfy the axioms of a topological space.
For example,
1) I give you the set ##X=\mathbb{R}##
2) I give the topology ##\mathcal{T} = \{\emptyset,\mathbb{R}\}##.
These two pieces of data specify a topological space, called an indiscrete space.
Other example:
1) I give you the set ##X=\{0,1\}##
2) I give the topology ##\mathcal{T} = \{\emptyset,\{0,1\},\{0\}\}##
These two pieces of data specify the Sierpinski topological space.
Obviously, not everything will be a topological space. For example
1) I give you ##X=\{0,1,2\}##
2) I give you ##\mathcal{T} = \{\emptyset, \{0,1,2\}, \{0,1\}, \{1,2\}\}##
These are again two pieces of data, but they do not specify a topological space because ##\mathcal{T}## does not satisfy the axioms. Indeed, ##\{0,1\}\cap\{1,2\} = \{1\}## is not in ##\mathcal{T}##.