• Introduction
Data science is the study of data; it may be structured or unstructured. It involve understanding, extracting values and visualize the data.
Various machine learning algorithms and statistical methods are used for this. It is the hottest topic of 21st century and the goal is to predict the information from the existing data.
Business intelligence (BI) is to make analysis and report with data, it is a subset of data science building predictive models help market to grow with great acceleration.
The following skills are required to be Data Scientist:
-
- Data Mining
- Data Analysis
- Data Visualization
- Statistics
- Machine learning
- Programming Language
1. Data Mining
-
- It is the technique of discovering patterns and extraction of useful information from the data.
- The other name of data mining is Knowledge Discovery of Data (KDD).
- For accurate model, we require more data.
• Stages of data mining:
-
-
Data Exploration:
-
This is the first stage of data mining, it consist of collecting data along with cleaning and transforming according to need of the problem. It can be done automatically as well as manually. For manual data exploration, by using queries and script in programming languages.
-
-
Modeling:
-
Data modeling is to apply the algorithms on the data and the goal is to choose the best data model based on the problem. Different model on the same data-set applied for choosing the best. Bagging, Boosting and Meta Learning are some popular techniques.
-
-
Deploying Model:
-
The final stage is the deployment of model that is the best in previous stage. It is important because the whole study based on this. Before deployment, we ensure the model is with the least noise.
2. Data Analysis
-
- It is the process of discovering useful results.
- Mined and cleaned data goes to analytic tools where it find patterns.
- In simpler term its analysis of past or future data.
- Data analyst use various techniques for analyzing data it can be done manually as well as automatically.
- Programming languages and analytic tools like R, python are used.
• Types of data analysis:
-
-
Text Analysis:
-
The analysis which is done on text data is called text analysis. It is a method used for converting data into important information which can be used in multiple industries. Sentimental analysis and lexical analysis are the part of text analysis. Text analysis help us to sort and rank the web pages.
-
-
Predictive Analysis:
-
Predictive analysis is the analysis of the unknown future result. It uses many techniques from machine learning and artificial intelligence. It combines the statistics with computational intelligence and result into the expected future values. Fraud detection and Risk management are some application of the predictive analysis.
3. Data Visualization
-
- It is the technique for visualizing the analyzed data.
- Large amount of data are very difficult to understand, that is why we use data visualization techniques as graphs and charts are easier to understand trends and pattern.
• Types of Data Visualization:
-
- Charts
- Tables
- Graphs
- Maps
-
- There are also many data visualization tools like QlikView Software and Fusion Charts, which help us to visualize the data without running any programmer.
- Python and R. can do manual data visualization.
4. Statistics
-
- It is the building block of all machine-learning algorithms.
- It help us to get deep and precise knowledge of data, which help us to study about the data.
- Without statistics, we can’t do machine learning and data science.
• Two categories of statistics:
-
-
Descriptive Statistics:
-
It provide information/description about the data. Data is categorized and organized based on the given parameter. It can be through the numerical value, table or by graphs
-
-
Inferential Statistics:
-
It predict the output based on the past data. The methods of inferential statistics is based on estimation of parameters and testing of hypotheses.
5. Machine learning
-
- It is a part of data science, the learning are on the data and its by computational machine.
• Algorithms are used for:
-
-
Regression:
-
It is a technique used to predict the dependent variable in a set of independent variable.
-
-
Classification:
-
It is a technique used for approximating a mapping function (f) from input variables (X) to discrete output variables (y)
-
-
Clustering:
-
It is a technique for dividing the population or data points into a number of groups such that data points in the same groups are more similar to other data points in the same group and dissimilar to the data points in other groups
6. Programming language
-
- Knowledge of programming language is must for writing the program to perform the art data science.
- There are many languages, which we can used. Python and R are most popular and used language
