Three directions of generalized data analysis

1. Data Analysis. Focuses on Analysis and Optimization of Small and Medium-sized websites, website maps, structure optimization, and SEO. Use third-party tools such as the open source analysis module (BIRT), CNZZ, and Google Analytics (GA ). Through analysis of website attribute data (such as pv, uv, proportion of new users, search term, bounce rate, bounce rate, access duration, and loyalty, optimize the website structure and content. This direction is more product-oriented and relies heavily on analytical experience and data sensitivity.
   Justin Cutroni, leader in website analysis, proficient in GA/GWO, as a blog Analytics Talk: http://cutroni.com/blog/
   SONG Xing, a Chinese representative, manages website analysis in China: Beijing/
2. Data Mining. Data Mining is mainly for decision-making. It is difficult to find unknown and intuitive conclusions from massive data. Such as content recommendation and relevance calculation. This work focuses more on internal data connections, data warehouse establishment, analysis system development, mining algorithm design, and even many times focus on processing raw data from ETL, therefore, there are high requirements on the computer level. Generally, data analysis is less extensive than data analysis, but more in-depth. In addition to programming languages such as Oracle, distributed computing Hadoop, C ++, Java, and Python, tools may also use third-party mining tools such as Weka.
   This direction is more technical, represented by Jeff Hammerbacher, former Facebook chief scientist, who participated in the compilation of "the beauty of Data". Some of the content is as follows:
   360doc.com/content...
   By reading "exploring the secrets inside receng", you can experience the charm of Data Mining:
   Ibm.com/develope...
   Ibm.com/develope...
   Ibm.com/develope...
3. Data Statistics. Focus on modeling and statistical analysis, establish reasonable models through mathematical knowledge such as probability, statistics, and discretization, and fully explore data content. For example, regression analysis is used to make full use of historical website data for evaluation, prediction, reverse prediction, and mining. Bayesian modeling is used for machine learning, clustering, and spam filtering. Common tools include SAS, R, and SPSS.
   This direction focuses more on mathematics, especially statistics. Hammerbacher, who graduated from Harvard mathematics, is also very strong in this regard. Data Statistics are not limited to the Internet, and are of great use in traditional industries, especially medical and financial fields.