**The issue of graphic analysis is to facilitate the study and conclusions.**

## Introduction

” *a good sketch is better than a long speech* ” Napoleon told us. There is the whole issue of graphic representation. It facilitates the sharing and understanding of information. The use of graphics is part of the logic of **visual management**.

## The design of a graphic

### 1. Identify the purpose of the graph

First step in the process, identify the objective of chart^{1}. Should it show a correlation, a trend…? Depending on its purpose depends on the type of chart.

### 2. Choose the graphic support

Make a choice and identify whether the graph will be used as part of a book, a PowerPoint…

### 3. Identify the type of chart

There are 5 graphics families. The choice is based on what you want to show^{2} :

**fluctuation or trend graph**: It shows an evolution of a parameter according to time (monthly sales…). We use curves or straights.**Decomposition graph**: It shows the proportions of the different elements in a set (sale by Product…). Most often, ”*Camembert*” graphics are used.**Distribution Graph**: It shows the distribution of an element relative to a set. These are the histogram charts.**Correlation Graph**: It shows the link or not that it exists between two sets of data. They’re clouds of dots.**Comparison graph**: It shows the difference between several datasets connected by the same parameter. For example, the financial result of the different companies in the same sector. These are stick, point or curve diagrams.

### 4. Collect Data

The collection is essential and must be carried out in a **rigorous manner**.

### 5. Build and validate the graph

Once the data is collected, build the graph.

Before you release the graph, you must validate it. This validation is in 2 criteria:

**Understanding:**Make sure that anyone who sees the graph understands the**same thing.****Objective:**to ensure that it meets the desired representation objective initially.**visual:**integrating pictograms, colors, effects…

## Fluctuation or trend graphs

Also called *time Series Plot*, fluctuation graphs represent data in chronological order. It allows to visualize changes and variations according to time and to understand the evolutions.

An evolution of this graph is the *Process Behavior chart*, used for **control charts**. This graph proposes to add 3 horizontal lines to improve the interpretation and have a more objective analysis than the *Time Series Plot* :

**UCL:**The maximum value accepted by the client**LCL:**The minimum value accepted by the customer**X cross:**average (UCL + LCL)/2

## Distribution charts

Even more than the others, the **Distribution charts** are an integral part of the statistics. This type of graph was used to represent populations in the first censuses.

The **Distribution charts** Allow to visualize the distribution of the same quantitative criterion with various values. Example: Age, weight, length… For example, the number of parts weighing between 10 and 15 kg, between 15 and 20 kg… This is useful in many cases of quality controls of parts, noise measurement…

### The box plot

There is another type of scatter chart known as the Box Plot (also called the *Japanese candlestick*, *Tukey diagram* or *moustache boxes*). A box plot indicates where the data center is and how the values are spread around:

- The box shows the 50% of the values in the middle.
- The vertical lines indicate the
**range**Values with the exception of Outliers. - An asterisk designates a Outliers (very large or very small).

This type of graph is used to visually compare data with input per attribute.

## Correlation graphs

Correlation graphs, also known as *Scatter Plot*, allow you to see the behavior of one parameter based on another. This type of diagram can be used to infer:

- If there is no correlation between 2 parameters
- The type of correlation: strong, weak, positive or negative.

The main use of correlation diagrams is when studying the causes of a problem. One can for example measure the relationship between 1 cause and one effect, or 1 cause and cause of the same cause or 2 causes between them. By measuring these parameters, we can define priorities. For example, if a cause has a very strong correlation with an effect, that cause must be a priority when solving the problem.

But more than that, the correlation diagram can be used for other purposes: for example, a correlation diagram tells us that the rate of delay in a meeting inflows the quality of these meetings. Thus, to set up an indicator, we prefer to use the delay rate, which is easier to measure, than the quality of the meetings.

## Decomposition graph

A decomposition graph is used to evaluate the composition of a set. In other words, compared to a quantity of 100%, a composition graph evaluates the proportion of each element.

Two types of graphic representation are found:

## Comparison graph

A comparison graph compares several set of values against one or more identical parameters. Two types of graphs are used:

**the radar graph:**The radar diagram was used for the first time in 1877 by Georg von Mayr

^{4}in his work “Die Gesetzmäßigkeit im Gesellschaftsleben”. The radar chart allows you to plot the values of each category along a separate axis that begins at the center of the graph and ends on the outer ring. It allows to compare several sets according to several criteria (no more than 10 in which case, the graph becomes illegible).

**the histogram chart:**when there is only one measurement parameter, a histogram chart is used.

The comparison graphs allow you to answer three questions:

- Which variables are dominant for a given criterion?
- Which variables are the most similar, are there clusters of variables?
- Are there any exceptions, variables that stand out?

## Source

1-W. Nerds (2014) – Economic statistics

2 – M. Ganuelas-Sinéchal, M. Vandercammen (2010) – Market Research: Methods and tools

3 – A. Schärlig, O. Blanc (2000) – Making the figures speak: descriptive statistics for the management Service

4 – Mr. Friendly (2008) – Milestones in the history of thematic cartography, statistical graphics and data visualization