Nowadays, large amount of data is generating day by day in every field like scientific research, medical, meteorological, climate, finance or marketing. To analyse the pentabytes of data across an extremely wide increasing wealth of weather variable, we need to build a platform for this, which is extremely flexible & scalable. Mapreduce is the most suitable & efficient parallel programming pattern for processing the big data analysis. There are so many frameworks have implemented to achieve high performance for the retrieval of information. This paper explains the analysis of weather forecasting data stored at NCDC and later processed by Mapreduce Framework. This paper also explains different frameworks of mapreduce for parallelizing computing heterogeneous & multi-core environments such as phoenix & phoenix++.