Machine learning in the heart of the decisional.

decision-making tool provides a basic way to understand and translate information by giving it a certain exploitable and usable value for better catch of decision.

From data collection to data exploration, decision-makers have products in their hands to translate their data more quickly into useful information for their activities in order to answer a number of questions. But these responses produced may not be enough.

Where is this insufficiency?

To properly locate the breach, it is first necessary to understand how the process of analysis and where it really ends. The BI mainly asks the questions of “what”. For example, what are the sales that have been made, what are the performances of the sales representatives, in Which region the sales were more profitable, which products are concerned, etc.?

When these numbers remain poor, decisions can be taken to correct deficiencies.

The analyses rather look for the causes, the ‘why’. Why is it that sales of a product decline over a given period of time? Is there a causal relationship between turnover in one region and the presence of others?

BI tools thus greatly facilitate data analysis but leave the final decision to the human being who does not have all the capabilities to provide an accurate future interpretation.

How to integrate computer intelligence into our decisions?

Many solutions and tools exist today on the BI but also in machine learning. In fact, the real challenge is to have a unified solution that provides a sophisticated interface like Power BI, Qlik, Tableau Desktop, google Data Studio and allows behind it to use intelligent algorithms.

Many companies are beginning to capitalize on this approach.

We must therefore to combine the decision-making tool with methods to model human intelligence by creating a certain intelligence behind the dashboards.

Detecting for example the products likely to interest a customer is not easy to achieve in the context of a conventional decisional analysis and that machine learning today can solve.

Tableau published in its version 10.2 TabPy, Tableau python server, an API that allows integration of Python code in Tableau. With TabPy, you can create fields calculated with Python codes. With all the current power of this language in the statistical calculation and the different libraries that exist today you can imagine everything you can make.

We will conduct an analysis of China’s GDP from 1960 to 2014 using data available in a csv file. our data set consists of two columns year and gross domestic product in US dollars as shown in the image below.

Leave a Reply

Your email address will not be published. Required fields are marked *