Tools of Data Science (G7, W7, M7)

This page is a memorandum of my idea.

I arranged G7, W7 and M7 as basic tools of Data Science . The original idea of these tools is from " Seven QC Tools and New Seven QC Tools " (Q7 and N7).

G7 and W7 are for beginners. M7 is for middle class users.

G7

"G" of G7 is "Graphs of Numbers". G7 is similar to Q7 .

Stratified Sampling is also useful in G7

Bar Plot : Analysis of size. ("Pareto Chart" is arranged bar garaph.)
Line Graph : Time series analysis. ("Control Chart" is one of the line chart.)
Circle Graph : Analysis of ratio
Histogram : Analysis of distribution of 1 variable. (One of the Bar Graph)
Box Plot : Analysis of distribution of 1 variable. (similar to stratified histogram)
Scatter Plot : Analysis of the relationship of two variables. Analysis of distribution of composed 2 variables. (To find outlier or to study small data set)
Heat Map : Analysis of the relationship of two variables. Analysis of distribution of composed 2 variables. (To study big data set)

"W" of W7 is "Analysis of Words". Most of W7 are Concept Analysis .

Association Diagram : Classification of idea. The method to collect Brainstorming
Fishbone diagram (Cause-and-Effect Diagram) : To collect reasons and results.
Tree Diagram : Similar to FMEA . The next step of Cause-and-Effect Diagram.
Association Diagram : Main part of Systems Thinking .
Matrix Diagram : Applications are QFD , Multi Dimensional Scaling and AHP.
Arrow Diagram : Planning method.
Flow Chart : Analysis of the process

Why-why Analysis is a basis of concept analysis

W7 analyze the struture of phenomena as levels or networks. The idea to think the structure is also useful in G7 and M7.

"M" of M7 is "Mathematical Analysis".

Error Analysis : Analysis of the qualtiy of the data
Statistical Value : Analysis of statistical value.
Hypothesis Testing for Diffrence of Average : Basic tool of Hypothesis Testing.
Regression Analysis : Include of the analysis Correlation .
Principal Component Analysis : In N7, called "Matrix Data Analysis".
Decision Tree : Application of Stratified Sampling
Linear Programming : To find the best in constraints.