Advances in ground-based characterization of transiting exoplanets using novel data analysis methods

Exoplanet characterisation is an essential aspect of contemporary planetary science aiming to understand the fundamental aspects governing planet formation and the origin of life. In this thesis, we present an array of advancements in ground-based spectroscopy of transiting exoplanets enabled through new data analysis techniques. We focus on low-resolution transit spectrophotometry observations commonly secured using multi-object spectrographs on ground-based medium to large telescopes. We introduce novel data-driven methods to correct for the telluric and instrumental noise and stellar variability that commonly contaminate the planetary signals in such observations. We show that our approach provides an alternative strategy to maintain the accuracy and precision of retrieved planetary signals, especially in cases when the conventional methods are suboptimal. We also show that our approach opens the doors to ground-based follow-up of exoplanets transiting very bright stars which have previously been difficult due to the non-availability of suitable comparison stars required for noise calibration in traditional methods. Eventually, we present a library of software tools implementing the methods presented in this thesis and demonstrate its application to a cumulative of nearly 200 hours of high signal-to-noise time series transit spectroscopy observations from the ground. In essence, this thesis contributes to the ongoing quest to push the technical limits of ground-based observations to the levels required for the characterization of terrestrial exoplanets in near future.