[en] For most of mankind's history, astronomy was performed on-ground in the optical energy range. It was only when space-based missions, built more than 50 years ago, detected photons with mind-boggling energies that the exploration of the violent Universe really began. These γ-ray photons still provide us with an unprecedented wealth of information for the most energetic processes taking place in the cosmos. Faithful to the olympic slogan ''higher, faster, further'', an increasing armada of γ-ray satellites was built and launched over the last couple of decades with Fermi being the youngest of its kind. In this thesis, I use data from the Gamma-Ray Burst Monitor (GBM) onboard the Fermi satellite. The focus of this work lies on three very different classes of high-energy astrophysical transients: Gamma-Ray Bursts (GRBs), solar flares and Soft Gamma Repeaters (SGRs). In Chapter 2, I present GRB 091024A, a burst of very long duration in γ-rays where optical data could be acquired well during its active phase. The optical light curve shows very intriguing features which I subsequently interpret as the so called ''optical flash'', a fundamental property of the ''fireball'' model. Although predicted by the latter model, only a handful of GRBs show such a behavior, making them interesting transients to study. Furthermore, I present the fundamental temporal and spectral properties of 47 GBM-detected GRBs with known redshifts. As GRBs explode at cosmological distances it is of uttermost importance to study them in their restframe to get a better understanding of their emission mechanisms. I confirm several correlations already found in the past together with an intriguing connection between redshift and the peak energy (E_peak) of GRBs. Although this correlation is heavily influenced by instrumental effects, it is not unexpected from other experimental results, giving it more credibility. Finally, I present the results of the search for untriggered GRBs in GBM data. This project focuses on GRBs which triggered Swift but not GBM although the GRBs came from positions above the horizon, with a favorable orientation to at least one GBM detector. The properties of these GRBs are then compared to the full sample of GBM GRBs published in the GBM spectral catalogue. Although designed mainly for GRB studies, GBM observes solar flares as well. In Chapter 3, I made use of the high temporal quality of GBM data to perform a detailed timing analysis of four solar flares. Contrary to recent claims in the literature, where quasiperiodic pulsations (QPPs) have been allegedly identified in the γ-ray data of solar flares, I did not find any statistical significant signatures of such QPPs. When red-noise, an intrinsic source component, is accounted for, most of the claimed QPPs fall below the threshold of a significant detection. Moreover, I developed a new background estimation method for solar flares, called SOBER² (SOlar Background Employing Relative Rates). This method uses the count rate of the complementary and shaded BGO detector as an a priori information to determine the background fluctuations for the Sun-facing BGO detector. Such a method is especially useful and beneficial for solar flares because they are usually of very long duration and the standard GBM background subtraction fails in such cases. Finally, in Chapter 4, I present the log-parabolic model which I used to fit the spectra of SGR bursts. Even though the spectra of the latter are usually and preferentially fit by a sum of two blackbodies in the literature, I show that the log-parabolic model fits the data as least as well as the double blackbody function. Additionally, the log-parabolic model is based on a strong underlying physical mechanism, i.e. second order Fermi acceleration, which gives it even more credibility.