Data Visualization

Data Science

The World can be Represented by All Kinds of Data

A lots of insights and significant facts and associations can be gleaned from the result of statistical analysis, but it is the data presentation and visualization combined that will give a greater meaning in conveying the hidden messages buried within these numbers.

Often the model of experiment could have many independent variables and dependent variables, and they may have been corelated to certain degrees. The univariate statistics often failed to address the significant relationship in the complex data set. Multivariate analysis involves observation and analysis of multiple random variables, while we analyze the joint behavior of more than one variable at a time and seek to better understand their contributions, as they could be influenced by multiple predictors, or even combination of several factors.

Multivariate analysis can clarify some of the data relationship and make a better explanation in coming up with a better model through hypothesis testing. There is a subtle distinction of statistical models between “multivariate analysis” having two or more dependent or outcome variables, and “multivariable analysis” referring to which there are multiple independent or response variables.

In multivariate statistics, we do not assume that the response variable is influenced by one factor alone, as it could be influenced by multiple predictors, or even combination of several factors. there are analysis methods to explore different features of the data include: principal cluster analysis, multidimensional scaling, latent class analysis, latent profile analysis, latent trait analysis, factor analysis, multiple regression analysis, discriminant analysis, etc.

These powerful statistical techniques are especially useful in biotech research to gain more meaningful insight while processing large volume of data associated with complexities of biochemistry.

Other multivariate statistical applications include: 1) Quality assurance across a range of industries such as food and beverage, paint, chemicals, pharmaceutical, energy, and telecommunications, etc., 2) Industrial process control and associated parameters optimization