THINK-AI DOCUMENTATION

French

A. Introduction

1.THINK-AI Overview

THINK-AI is a graph generation platform that combines artificial intelligence with advanced dataset visualization. It transforms raw data into actionable information, enabling users to explore complex patterns, trends and correlations. Whether analyzing financial data, scientific research or business data, THINK-AI goes beyond traditional graphs and tables to reveal the insights hidden in the data.
The platform also enables the creation of survey questionnaires and polls for collecting data from a variety of stakeholders, such as respondents, prospects and customers. We designed THINK-AI with the aim of making data exploration and management more intuitive and accessible to all those seeking to fully exploit the value of their data. Each user with an account can customize a dashboard, where generated graphs and reports can be saved and managed.
This guide gives a detailed overview, with illustrations, of the platform’s various functionalities and how to make the most of them.

2. Target audience

THINK-AI is aimed at a wide range of users, from data professionals and researchers to financial analysts and business managers looking to maximize the value of their data. Accessible to all, THINK-AI is designed for users of all levels, from novices to experts in data management and analysis.

3. Key features

THINK-AI’s functionalities fall into three main categories:

■ Data visualization: THINK-AI enables you to create interactive visual representations from complex data, making it easier to analyze and understand information,

■ Mathematical model generation: The platform integrates artificial intelligence algorithms for the creation and analysis of predictive and explanatory models, optimizing data-driven decision-making,

■ Surveys and surveys: THINK-AI offers tools for the creation and management of questionnaires and surveys, enabling the collection, analysis and interpretation of qualitative and quantitative data from various target populations.

4. Official website

You can access the THINK-AI platform via the following official link: https://www.think-ai.tech/. This link is accessible from all browsers.

B.Connecting to THINK-AI

To connect to THINK-AI for the first time, you need to create a user account by going to the Create an account menu on the home page. After completing the registration form, validate your registration by clicking on Register. Once registration is complete, you will be able to log in using the identifiers created at this stage.
We recommend that you keep your login details in a safe place, as they will be required for all future connections to the platform.
The images below illustrate the process of creating an account on THINK-AI:

■ Go to the Create an account menu on the home page,

User account creation menu

■ Fill in the account creation form with the requested information, then click on Register to validate the registration,

Registration form

■ Return to the Log in menu, enter the user name and password you created earlier, then click on the Log in button.

Connecting to the THINK-AI platform

C. User interface

THINK-AI’s user interface is structured around two main menus:

■ The Summary menu, which lets you manage and manipulate datasets. This menu is subdivided into several key submenus:

★★ Dashboard : provides an overview of generated data and graphs, making it easy to navigate and quickly analyze important information,
★★ Data : enables management of datasets used for visualization and modeling,
★★ Contact : provides the information needed to contact the platform’s technical or sales support,
★★ Documentation : contains detailed guides for using the platform and its features,
★★ Pricing : details the platform’s various pricing offers, according to users’ needs.

Summary menu and submenus

■ The Poll and Survey menu is dedicated to the creation of questionnaires for surveys and polls, enabling responses to be collected from participants. This menu offers an intuitive interface for designing customized questions and distributing surveys to respondents.

Poll and survey menu

D.Summary tab: Its menus and functions

1. Data menu

The Data menu is accessible from the left-hand sidebar, under the Summary tab. It allows you to manage and manipulate datasets used for analysis and visualization on the THINK-AI platform.

“Data” menu

The Data menu provides functions for the efficient manipulation of data sets. The various options available in this menu are described in the following paragraphs.

1.1 Import data sets

This feature allows you to import datasets from various sources into the THINK-AI platform. Supported dataset formats include:

■ Excel (.xls, .xlsx),
■ Text files (.csv, .txt, .json),
■ Relational Database Management Systems (RDBMS) (.sql).

a. Importing a .csv data set

The steps for importing a .csv dataset into THINK-AI are as follows:

■ Navigate to Summary > Data > Data connection,
■ Select as Data source csv,
■ Click on Browse to choose the file to import (in the example below, we select the nutrition2.csv data set),
■ Choose the appropriate Column separator (in this example, we use the comma, as our data set is comma-delimited),
■ Check Header if you want the first row of the data set to be considered the column header (this option is enabled in the example below),
■ Check or uncheck Delete spaces as required (this option is enabled in the example below),
■ Check or uncheck Omit ’' as required (this option is unchecked in the example below),
■ Check or uncheck Ignore quote “ as required (this option is disabled in the example below),
■ Under Missing value, select the treatment to be applied to missing data,
■ Under Comment, select the treatment to be applied to comments,
■ Finally, click on Import/Update to validate the import of the .csv data set.

Importing a .csv data set

Once the dataset has been imported into THINK-AI, it can be saved on the platform by clicking on the Save file button. This action automatically adds it to the list of uploaded data, thus avoiding multiple imports of the same dataset.

Saving a dataset on THINK-AI

b. Import an .xls data set

The steps for importing an .xls data set into THINK-AI are as follows:

■ Navigate to Summary > Data > Data connection,
■ Select Data source as xls,
■ Click on Browse to choose the data set to import (in the example below, we select the data set Flowers-Iris.xls),
■ Check Headers if the data set contains headers,
■ In Sheets, choose the sheet to export if the data set contains several sheets (in this example, our data set contains a single sheet named data),
■ In Data name, enter the desired name for the sheet to be loaded,
■ In Cell range, specify the range of cells to be imported,
■ In Ignore rows, specify the row number to be ignored if necessary,
■ In Missing value, specify the values considered to be missing,
■ Check First row as column name if necessary,
■ Finally, click on Import/Update to validate the import of the .xls data set.

Importing an .xls data set

c. Import an .xlsx data set

The import process for .xlsx data sets is identical to that for .xls data sets (see section Importing an .xls dataset).

d. Import a .txt data set

The process for importing .txt data sets is identical to that for .csv data sets (see section Importing a .csv dataset).

e. Import a .json data set

The steps for importing a .json dataset into THINK-AI are as follows :

■ Navigate to Summary > Data > Data Connection,
■ Select as Data source json,
■ Click on Browse to select the dataset to import (in the example below, we select the sample_data.json file),
■ Check Headers if the dataset contains headers,
■ Finally, click on Import/Update to validate the import.

Example of importing a .json data set

f. Import a .sql dataset

The steps for importing a .sql dataset into THINK-AI are as follows :

■ Navigate to Summary > Data > Data connection,
■ Select sql as Data source,
■ In Drive, choose the corresponding drive from those available in the drop-down list,
■ In Database name, specify the name of the source database (containing the data to be read),
■ In Database server, specify the name of the database server,
■ In Connection ID, enter the user name for accessing the database,
■ In Password, enter the database connection password,
■ Under Data name, specify the name to be given to the data to be read,
■ List of old data displays any old data read,
■ Under Data description, specify a description of the data,
■ Under SQL code, write the SQL code to be executed to extract the data from the source database,
■ Finally, click on the Execute/Refresh button to execute the SQL code in the source database.

Importing an .sql dataset

1.2 Transforming a data set

Dataset transformation refers to the processes by which the dataset is modified, enriched or converted into a format more suitable for analysis or visualization. To transform a dataset in THINK-AI, it must first be imported (see section Import data) and the following steps followed:

■ Click on the Data transformation menu under Data,
■ Under Column names, the names of the various columns in the imported dataset appear,
■ In Data type, the type of each column is displayed and can be modified as required,
■ Check Delete missing values if the column concerned contains missing values,
■ In Replace missing values, indicate the value to be used to replace missing values,
■ Under Rename column, enter the desired new name for the column,
■ Click on the eye icon under Summary to view a preview of a specific column,
■ Finally, click on the Update button to validate the transformations performed.

N.B.: The Search box at top right allows you to quickly find a column by entering its name.

Transforming a dataset

The following figure shows the Summary of the Species column of the Flowers-Iris dataset, which were imported according to the Data transformation steps mentioned above. This summary shows, in the form of a bar chart, the frequency of appearance of each species (here, the species are SETOSA, VERSICOLOR and VIRGINICA) in the Species column.

Summary of the SPECIES column in the Flowers-Iris dataset

1.3 Summarizing a data set

After importing and/or transforming a dataset, click on the Data summary menu to obtain descriptive statistics for each column, according to their respective type. By selecting a variable under Variable in row and another under Variable in column, then specifying the Graph type, it is possible to visualize the relationships between the two variables chosen.

The figure below shows a summary of the Flowers-Iris dataset. THINK-AI also allows you to download this summary in several formats, as shown in the figure.

Flowers-Iris data summary

1.4 Running an R script on a dataset

THINK-AI allows you to write and run R scripts on imported and/or transformed data. To do this, simply go to the R menu and follow the steps below:

■ Write the R script in the dedicated console (left),
■ Click on the Run button to execute the R script (execution results appear on the right),
■ Click on the Save code button if you wish to keep the script you have written,
■ Finally, if necessary, click on Clear console to delete the contents of the console.

Running an R script on a data set

1.5 Displaying a data set

Once you have imported a dataset and performed the necessary transformations, you can display it by accessing the Display menu. In the following paragraphs, we’ll explore the options available under this menu.

a. Hide columns of a data set

To hide certain columns in a dataset, proceed as follows :

■ Access the Column visibility submenu, then tick the columns you wish to hide,
■ Uncheck the columns you no longer wish to display.
To redisplay hidden columns, simply tick them again.

In the following example, we choose to hide the sepal_width and petal_width columns of the Flowers-Iris dataset.

Hide columns in the Flowers-Iris data set

b. Transforming columns in a data set

The Transform submenu lets you perform the following transformations on the columns of a dataset:

■ Rename column,
■ Descriptive statistics for the column,
■ Change column type,
■ Format column,
■ Create a new column,
■ Replace values,
■ Delete column.

Let’s examine these transformations one by one.

Transform submenu options

b.1. Rename column

This function allows you to change the name of a column within a dataset. To do so, simply enter the new name in the appropriate field, then validate the operation.

In the following example, the OBS column in the Flowers-Iris dataset is renamed to OBJET.

Rename the OBS column in Fleurs-Iris

b.2. Univariate and multivariate column analysis

This function allows you to generate statistical analyses for a column or set of columns in a dataset. It is divided into two parts: univariate and multivariate analysis.

■ Univariate analysis
Univariate analysis provides a comprehensive statistical summary of the values in a single column. This includes descriptive measures such as :

★★ Central trend: Mean, median, mode,
★★ Dispersion: Standard deviation, variance, amplitude, quartiles,
★★ Distribution: Histograms or density diagrams (depending on data type).
These indicators provide an overview of the main characteristics of the column under study.

For example, in the image below, the univariate analysis of the OBS column of Flowers-Iris shows a summary of values with statistics calculated for each element.

Univariate statistics for the OBS column in the Flowers-Iris dataset

■ Bivariate analysis
Bivariate analysis enables the study of relationships between two variables through pivot tables and corresponding visualizations. This enables us to understand the correlation or interaction between two columns in a data set.

★★ Creating a pivot table
To generate a pivot table, follow these steps:

★★ ★ select variable for columns under Choosing variables in columns,
★★ ★ select variable for rows under Choosing variables in rows,
★★ ★ Define variable to aggregate under Value to aggregate,
★★ ★ Choose Aggregation function (sum, average, count, etc.),
★★ ★ Use Filter field to add filter conditions or criteria,
★★ ★ Click Generate dynamic table to produce results.

The illustration below shows the process of creating a pivot table on the Flowers-Iris data set.

Creating a pivot table

★★ Application of filters on pivot tables
To refine your analysis, you can add conditional filters to pivot tables. The steps involved in creating a filter are :

★★ ★ Select the column to be filtered under Filter column,
★★ ★ Choose the appropriate Filter function (e.g. equal to, greater than),
★★ ★ Enter formulas or expressions in the editing area,
★★ ★ Validate by clicking on Validate or cancel with Cancel.

The image below shows the process of creating a filter on a pivot table.

Creating a pivot table filter

b.3.Changing a column’s data type

To change the data type of a column, select the desired new type from the drop-down list below the Change type option. This feature allows you to adapt columns to specific analysis or visualization requirements.

An illustration of this process is shown below.

Changing the data type of a column

b.4.Modifying column format

To change the format of a column, access the Column formatting option and select the desired format type from the available choices. Next, enter any additional information required in the appropriate fields. Finally, click on the button at the bottom of the window to Validate column formatting.

This procedure is illustrated below.

Modifying column format

b.5.Substituting values in a column

To substitute values in a column, select the Replace values option and follow the steps below:

■ In the Values to replace field, specify the values targeted for replacement,
■ Under Replacement values, specify the new values to be entered,
■ Finally, click on the Validate value replacement button at the bottom of the window to confirm your changes.

The above steps are illustrated below.

Substituting values in a column

b.6.Creating a new column

This feature lets you create a new column from a function applied to a source column, or recode an existing column.

■ To Create a new column, follow the steps below:

★★ In Choice functions to apply to source column, select the desired function,
★★ In Source column name, specify the name of the original column,
★★ In New column name, specify the name of the new column,
★★ If a specific condition must be met before the new column is created, activate the If condition option and fill in the necessary information (optional),
★★ Finally, click on the Validate new column creation button to confirm your changes.

Creating a new column

■ To Re-code an existing column, proceed as follows:

★★ In Source column name, specify the name of the column to be modified,
★★ In New column name, specify the name of the new column,
★★ Click on Category coding, fill in the required fields, then click on the Validate button to confirm the conditions for re-coding the old column,
★★ Finally, click on the Validate button to finalize the creation of the new column..

Re-coding an existing column

b.7.Deleting columns

To delete a column, select Delete Column. A confirmation message appears on the screen; click OK to validate the deletion, or Cancel to refuse the operation.

Here’s an illustration of this feature.

Deleting a column

c. Save the state of a data set

The Status Backup feature allows you to capture and store different versions or configurations of a dataset. This makes it possible to restore a previous state at any time, for rigorous tracking of changes.
To do this, access the Save Status menu and follow the steps below:

■ Select Create state,
■ Assign a name to the report using the Rename field,
■ Confirm (or cancel) the creation of the state using the Rename (or Cancel) button.

Illustration below.

Creating a state of a data set

Once created, the data set’s state can be edited (via the edit button, framed in red in the figure below), updated (button framed in green), or deleted (button framed in blue). These actions enable flexible and controlled management of saved states.

Options for modifying a saved state

↓