Big Data Elements

Big Data Elements

1. Business analytics

   • Analytics
     • It is a process that involves the use of elements such as statistical techniques, information and research methodologies to explore, visualize, discover and communicate patterns or trends in the data.
   • Types of analysis
     • Descriptive
       • Statistical techniques are applied simple that describe what contains a set of data.
     • Predictive
       • Advanced statistical methods are used to identify predictive variables and build models predictive to identify trends and relationships which cannot be obtained by descriptive analysis.
     • Prescriptivo
       • Science methodologies are used decision, management science and research of operations to make the best use of the assignable resources.
   • Definition
     • It is a process that begins with data collection related to the business and performs a sequential application of main descriptive analytical components, predictive and prescriptive, the result of which supports decision-making business decisions.
   • Big Data Analytics
     • Allows organizations carry out processes such as:
       • Predictive analytics
       • Language processing natural
       • Image analysis
       • Statistical techniques such as modeling discrete choices and optimization math
       • Combine unstructured content and analyze it with your BI.
     • It allows exploring new methods that have the ability to detect correlations between amounts of data available.
     • Disciplines that support big data analysis
       • Statistics
         • Statistics provides the theory to test hypotheses about various knowledge from the data
       • Machine Learning
         • Machine learning (ML) allows the implementation of data mining based learning agents
       • Data Mining
         • Allows you to extract and acquire useful knowledge from data
       • Text Mining
         • Refers to the technique that automates the processing of large volumes of textual content for extract key trends and identify statistically the different themes that arise.
       • Management systems databases
         • These systems include relational DBMSs, NoSQL databases and databases big data, like Hadoop.
       • Flow management systems of data
         • handle transient flows, including queries continuous, and at the same time can handle data with very high intake rates.

2. Data

   • Definition
     • It refers to the basic and fundamental unit that is used in the analysis of a phenomenon, is the crudest element that we have and that will be consumed by the analysis process.
   • Characteristics
     • Origin
       • Source and reference of acquisition of each value as well as the means to obtain it.
     • Tipe
       • How the value is presented is text, number, position, image, sound, etc.
     • Granularity
       • Level of detail or specificity of the data in a set of data, indicates how much detailed information the data contains individual.
     • Structure
       • Form of data representation, whether value simple, document or multimedia.
     • Uncertainty/error
       • Level of certainty about the correctness or precision of the data or measurement error
     • Quality
       • Joint assessment of the attributes of origin, reliability and Data usability is an approach to evaluate and measure in What measure is appropriate to use the data.
   • types of data in digital age
     • Databases
       • They are the main source of data, there are technologies such as hadoop, SQL that allow the data recovery and storage.
     • Raw data
       • Complex data that requires preprocessing specific to make them manipulable by someone algortimo.
     • Text
       • Refers to any text written in natural language human iincludes text types such as; books, articles, news, etc.
     • Multimedia content
       • Audio, images and videos are useful for capturing data and They present themselves as a challenge for data scientists.

3. Statistical Methods

   • Algorithms
     • They use big data to process data and generate models that learn from data to predict behavior futures and perform smart tasks.
     • qualification
       • Preprocessing
         • The data is received in its pure state from different sources, the information is identified and extracted. information, then the data is cleaned and the values ​​required in the data, if empty fields are found, the data imputation technique is used.
       • Analysis
         • The creation of models is carried out to support the desired tasks, where trains the model to evaluate effectiveness before it is put into production.
       • Visualization
         • Scientific methods are used to transmit the information obtained in the analysis of data by generating graphs and diagrams to achieve visualization effective.
   • Methods
     • Media
       • Average mean is the initial technique to perform statistical analysis, it allows you to see the data While determining the general trend of the data collection, the mean determines the center of the data under consideration.
     • Standard deviation
       • It allows measuring how widely the data is distributed with respect to the mean, frequently used to analyze the spread of data points whether they are grouped or No, for example, a high standard deviation indicates that the data is widely dispersed. of the average
     • Regression
       • It studies the connection between an independent variable and a dependent variable, i.e. how one variable can influence another or how changes in one variable result in changes in another. Regression graphs and diagrams use lines to indicate trends during a given period, and the strengths or weaknesses of the correlations between variables.
     • Hypothesis testing
       • It allows measuring how widely the data is distributed with respect to the mean, frequently used to analyze the spread of data points whether they are grouped or No, for example, a high standard deviation indicates that the data is widely dispersed. of the average
     • Sample size
       • It consists of examining a sample or data that is smaller than the data set, You must choose a suitable sample size otherwise the result will not be reliable, and choosing an excessive sample size can cause loss of time and money.