Rapport de projet Business intelligence

 2015-2016

Housing in Tunisia

in cooperation with CNS

Plan

Plan .......................................................................................................................................................... 2

Introduction ............................................................................................................................................. 3

Project context ........................................................................................................................................ 4

Modeling ................................................................................................................................................. 5

Data Integration ...................................................................................................................................... 7

Analysis and reporting ............................................................................................................................. 8

Big Data ................................................................................................................................................. 12

Data mining ........................................................................................................................................... 16

Figure 1 Data warehouse model 6

Figure 2 Extract Transform Load (ETL) 7

Figure 3 Number of sales per year 8

Figure 4 Cities classification per number of sales 8

Figure 5 Cities classification per square meter price 9

Figure 6 Regions classification per number of sales 9

Figure 7 Sales global Dashboard 10

Figure 8 global Dashboard 10

Figure 9 Dynamic map of sales information 11

Figure 10 Big Data: Facebook 13

Figure 11 Big Data: Twitter 14

Figure 12 Correlation between variables 16

Figure 13 m² price evolution 18

Figure 14 Correlation between sales and m² price 18

Figure 15 Sales number in function of time 19

Figure 16 m² price in function of time 20

Figure 17 Essour activities distribution 20

Figure 18Essour word cloud 21 3

Introduction

Nowadays, data is important to make decisions, but we have a huge amount of raw data that could alter the visibility in order to reach our goals, and take the proper decisions.

Data are now woven into every sector and function in the global economy, and, like other essential factors of production such as hard assets and human capital, much of modern economic activity simply could not take place without them.

Business Intelligence could be the alternative to give sense to a big quantity of raw plain data to make meaningful and useful information.

For example, it could be a solution to have an idea about the evolution of housing in Tunisia which faces today a lot of challenges. 4

Project context

This project is elaborated in cooperation with the National Statistics Council (CNS) which was created within the law number 32 of April 13th, 1999 relative to National Statistics System.

Their main activities are:

 Coordination of activities between the different structures and organization that are in charge of statistics.

 Publishing and Broadcasting of the generated statistical information.

 Organize consultation between producers and users of statistical information.

 Collect data from households, enterprises, administrations…

 Define concepts nomenclatures and standards to be adopted internationally.

From the CNS we obtained Row data about the real estate market in Tunisia (prices, contract types, regions…)

Using this data, we tried to analyze the evolution of housing in Tunisia, visualize and study the general behavior of the housing market, predict the prices evolution during the next years in order to define and overcome the multiple challenges that this sector faces like:

 Economical evolution

5

Economic instability especially for the post revolution era.

 Regional inequity

The state of housing is globally better in costal urban regions than any others.

 Social challenge

Within a context of high inflation and unemployment.

Substandard housing struggle.

Modeling

1. Client requirements:

Based on the row data provided by the client he wants to:

 Follow the evolution of the real estate market

 Explain the evolution of the price through time (eventual increases), property type and place (city and region).

 Link important peaks and falls to remarkable events (political, economical… ).

2. Our solution:

In response to these requirements we elaborated the model of our data ware house which makes the information:

 Easily accessible

 Consistent

 Adaptive and resilient to change

 Serve as the foundation for improved decision making

We are going to focus on the sales of the real estates and the different characteristics that it depends on it like region, time … 6

Figure 1 Data warehouse model 7

Data Integration

During this phase we faced a lot of challenges due to the poor and corrupted data, unusable fields, Lack of documentation and the Lack of analysis axes…

At the end using this model, we’ve been able to organize the row data that we have and transform it into meaningful homogenous and centralized data.

Now that we have useful, organized data we continued to analysis and reporting phase.

Figure 2 Extract Transform Load (ETL)

Tools:

 Talend: Data integration software.

 PostgresSQL: Open source database.

8

Analysis and reporting

Using analysis and reporting tools we were able to create these dashboards that responds to the client different requirements and will help them eventually in the decision making.

...