Big data: concepts, applications, and implications

Big Data

Nowadays, enterprises are employing statisticians when carrying out sophisticated data analysis. This is caused by the increased affordability in data acquisition and data storage among large scale and small-scale enterprises. This article focuses on highlighting the emerging Magnetic, Agile, Deep (MAD) data analysis. This is figured out as a shift from traditional enterprise data intelligence. The article presents its design philosophy as well as experience and techniques that portray MAD as one of the biggest advertising networks for the interactive media. Moreover, data parallel algorithms are presented for sophisticated techniques putting more emphasis on density methods. Another crucial reflection of the article is the database system features, which facilitate agile design and flexible construction of the algorithm (Cohen, Dolan, & Dunlap 2009).

The article has included the knowledge from some prior study. For instance, the standard business practices applied in the large scale data analysis revolves around the notion of the Enterprise Data Warehouse. This is queried using the Business Intelligence software. This software is used to produce reports and interactive interfaces, which summarize data through the elementary aggregation functions (Cohen, Dolan, & Dunlap 2009). This topic was useful for academic research and industrial development in the 1990's. As the orthodox approach continues today in many environments, there are crucial factors, which are pushing towards a different philosophy.

The article has various findings and points. First, storage has become cheap. This is because few subgroups subgroup can construct an isolated database in any enterprise. In any department, one to two orders with magnitude can be paid than that which is not coordinated by management. Moreover, the traditional approaches of EDW repel the new data sources. This discourages their incorporation until full integration and cleansing is carefully done. The current data analysis involves the highly complicated statistical methods, which go beyond the drill-downs of traditional BI. The article illustrates the techniques and experiences developed during the construction of the MAD (Cohen, Dolan, & Dunlap 2009).

The study presented in the article has a lot of significance. First, it discusses the database design methodology, which in turn focuses on facilitating an agile and an organized approach to data analysis. Moreover, the study emphasizes on critical database system characteristics, which foster agile design and emphasize on flexibility in algorithm development. This includes high performance, data ingress as well as the heterogeneous storage facilities. In the article, the need for the unified system is emphasized. This is because they integrate and embrace an extended variety of data insensitive programming methodologies. In this case, the analysts come from different aspects of life. Moreover, this entails moving beyond the religious debates concerned about the benefits of SQL over Map Reduce. The researchers argued that many data storage formats and data sources should be combined using the parallel dataflow engine (Cohen, Dolan, & Dunlap 2009).