With the avalanche of data and increasing sky coverage of surveys, we strive to improve our "machinery" in Cosmology Analysis. In this talk, I will concentrate only on the large scale clustering analysis. As in all cosmology analysis, we start with the simulation of the mock catalogs, we applied machine learning (ML) techniques in predicting the number of galaxies that should occupy a halo given its properties. The ML techniques here distinguish themselves from the traditional halo occupation distribution (HOD) modeling as they do not assume a prescribed relationship between halo properties and number of galaxies. We will then turn our attention to the data themselves. As described in Padmanabhan et al. 2012, we see the application of reconstruction techniques improved the measurement of Baryon Acoustic Oscillations by a factor which is equivalent to increasing the surveyed volume by a factor of 3. In this talk, we will describe our recent effort in improving the technique in actual data, based on theoretical and numerical efforts by Tassev & Zaldarriaga 2012.