BIG DATA STRATEGY AND CHALLENGES

1.5.1 STRATEGY

The strategy is composed of five phases that involve different activities [7]:

• • The hardware analysis is required for installing the software and the data to be analyzed, with the recommendation of a data server with a large storage capacity.
• • The selection of the company’s processes that will be analyzed can be customer sales processes, production data, equipment failures, among others; this process selection collects the necessary information and data that will be the raw material for the subsequent activities.
• • The installation and configuration of the Hadoop platform are distributed data processing, as well as the software to support the Hadoop system.
• • The extraction, transformation, and loading (ETL) activities with analysis services.
• • The big data analytics, tools for analyzing reports (reporting), queries, and visualization (dashboards) will lead to data analytics.
• 1.5.2 CHALLENGES

The application of mass data has considerable benefits for individuals and

society, but it also raises serious concerns about its potential impact on the dignity, rights, and freedom of the persons concerned, including their right to privacy.

These risks and challenges have already been the subject of multiple analyses by data protection specialists around the world. We can identify the two following concerns [14]:

• 1. Lack of Transparency: As the complexity of data processing increases, organizations often claim secrecy about how data are processed for reasons of commercial confidentiality. As in 2014, the White House Report noted, “some of the most important challenges revealed by this review are how massive data analysis can create a decision-making environment so opaque that individual autonomy disappears into an impenetrable set of algorithms” [15]. Unless natural persons receive appropriate information and have adequate control, individuals cannot exercise effective control over their data and give informed consent when required. This is particularly true for the precise future purposes of any secondary use of the data that may not be known at the time of data collection. In this case, the controllers may not be able or willing to explain to the data subjects precisely what will happen to then data and obtain then consent, if necessaiy.
• 2. The Information Imbalance: Between the organizations holding the data and the data subjects whose data they process is likely to increase with the development of applications based on massive data [14].

It should be mentioned that big data also have several weaknesses, such as:

• • Detecting data judged abusive or earlier all data that will not follow a dominant statistical model, and we systematically remove any data contrary to the dominant statistical law.
• • The absence of the quality of being reliable about results, the big data has the farcical tendency indeed to inspire and process a maximum of data, but without making a quantitative sorting. Here we can mention the famous example of the giant of the web Google, in 2011 to adopt a project to call Google flu trend to make a study on the evolution and the appearance of the influenza epidemic using its algorithm developed; they can collect data input on search engines as keywords.

for example, cough, flu, fever. However, the result was ambiguous and overestimated.

• The difficulty of processing what has not been detected and anticipated, and this makes us a tool that little performs at the novelty and breakdown.