During the last few decades, condensed matter physics has witnessed a deep refoundation of its paradigms, through the discovery of many systems that the usual symmety classification à la Landau cannot handle properly. Although the first major breaktroughs were realized at the time of discovery of integer and fractional quantum Hall effects, only recently physicists have agreed that these peculiar phases of matter require neither a magnetic field nor low temperature. Those new states of matter cannot be caracterized by the geometric aspects of the model but rather by topological ones. The precise shape of the electronic spectrum is no longer relevant, but only particular features are, such as the presence or the absence of a gap. Similarly to the Landau classification scheme, one can achieve a construction through extensive use of topological groups. This is the realm of algebraic topology. Related generated topological invariants can hold a classification of non-trivial topological states, as well as of the accompanying transitions. This thesis focusses on peculiar topological features of two-dimesnsional electronic band structures. After a technical introduction to the underlying formalism, the first chapter is devoted to local topology, that is for a restricted piece of the first Brillouin zone, of band crossing points, also known as Dirac points. Special care is taken to classify these points and related transitions. The next chapter sheds some light on a particularly efficent way of measuring topology for two-dimensional electrons. This is achieved through measurements of Landau levels that are generated by a magnetic field applied perpendicular to a plane. Dirac points then generate zero Landau levels that are topologically stable, i.e. almost not influenced by perturbations at all. Distinctions between low and high magnetic fields will prove to be relevant, although very system-dependant. Through the several models studied, we particularly stress out the importance of the topological tool for condensed matter physics, past present... and future.
