Descriptive Analytics tools and techniques
Descriptive analytics is the conventional form of Business Intelligence and Data analysis which seeks to provide a description or summary view of data in the form of graphs and numbers in common layman terms. Descriptive Analytics uses Data mining and Data aggregation to extract this information. Descriptive analytics present past data which is easy to understand by the business organisation. Read more What are the descriptive analytics tools, methods, and techniques?
Descriptive analytics helps to describe and present data in a simple format that can be easily understood by the business organisation. Data Analyst is really concerned to investigate cause and effect relationships. After conducting a successful survey, they draw an important conclusion which can be easily implemented in the organisation. Some common methods involved in Descriptive analytics are case studies, observation and surveys.
Descriptive analytics includes both numerical (mean, median, mode, standard deviation, variance etc) and Graphical methods (histogram, regression, correlation, pie chart, box plot etc) which helps us to summarise the data and extract the useful information like central tendencies, dispersion etc. These data can help us to describe the association between the several variables.
In order to estimate the right descriptive analytics tools and techniques, we need to consider the form and type of data and the number of variables associated with them to have an overall analysis. Based on this information, we have created a complete overview that will help you to describe which tools to use accordingly to your preferred situation.
We can broadly classify these set of data types into three main categories:
- Quantitative: These type of data set are mainly used to describe the quantities of the object of interest. They mostly show the s=results in the number format. Some of the classical examples are- average run scored by the player in one day match, total aggregate scored by the student in any exam, the weight of the boy and the girls of the classroom etc.
- Qualitative: These type of data set are mainly used to describe the qualities of the objects of interest. They mostly show results in the form of categories, which are also referred to as levels or modalities. Some of the classical examples are- The gender of the IT professionals working in software companies, the type of slipper produced by industry etc. These can be measured either in male, female or transgender or rubber or leather.
- Mixed: These types of data set are mainly used to show Qualitative and Quantitative data. They are expressed in the form of numerical values and categories. The most common example of mixed data is a number of people getting registered for the census of India. this category includes the total sum of people and their gender.
?Here are the tools widely used for Descriptive Statistics. These tools can handle either the description of one of the descriptions of the relationship between two or more variables.
Qualitative Analysis
One variable (univariate analysis)
- One variable can be used to estimate the frequency distribution of any given data. The numerical tool used is the frequency table. The histogram is the most common type of graphical tool used to measure frequency distribution. Example– How many people per age attended the guest lecture organised in the college? ( Here the investigated variables AGE is in?a qualitative form.)
- One variable technique is used to measure the central tendency of one sample. The numerical tools used in measurements of central tendency are mean, median and mode. Box plot Scattergram Strip plot is the most common type of graphical tool used to measure the central tendency. Example– What is the average weight of the class.
- ?One variable is used to measure the dispersion of one sample from the data. The numerical tools used are range, standard deviation, variance, coefficient of variation and quartiles. Box plot Scattergram Strip plot is the most common type of graphical tool used to measure the dispersion of the sample from the data. Example-How widely or narrowly is the salary dispersed from the mean salary of the industry.
- One variable is used to characterise the shape of a distribution. the numerical tools used for the measurement are skewness and kurtosis coefficient. The histogram is the most common type of graphical tool used to measure the shape of the distribution. Example– Is the teacher?s wage distribution in the company symmetric?
- One variable is used to visually control the data whether they follow a given distribution or not. There are no such numerical tools discovered to measure such distribution. The probability plot is used as a graphical tool to measure such distribution. Example-?What are the theoretical percentage of students who obtained better marks than a given threshold.
- One variable is used to measure the position of a value within a sample. The numerical tools used for the measurement of these types of data are Quartiles or percentages. Box plot is the most common type of graphical tool used to measure the position of value. Example– What data point can be used to split the sample into 90% of low values and 5% of high values?
- ?One variable is used to detect extreme values. Box plot is the most common type of graphical method used to detect extreme values. Example– Is the weight of 55kg an extreme value in this group of students?
Two variables (bivariate analysis)
Two variable is widely used to describe the association between two variables. The numerical tools used for the measurement is Correlation coefficients. Correlation Map Scatterplot are used as graphical tools to measure the association between the variables. Example- Do viruses increase or decrease during the monsoon season.
Several Variables
Several variables are used to describe the association between multiple variables. The numerical tools used for the measurement is Correlation coefficients. Motion Charts ( up to 3 variables to describe over time), Scatterplot 3-D ( up to 3 variable to describe)are used as the graphical tools to measure the association between the multiple variables. Example- What is the size of the population and the life expectancy over the last 20 years in this country?
Two matrices of several variables
This variable is used to describe the association between the two matrices. The numerical tools used for the measurement between such association are RV Coefficient. Example- Does the discovery of a series of medicine developed by Sun Pharma differ from the one from Mankind?
Qualitative Analysis
One variable( univariate analysis)
- One variable qualitative analysis helps to compute the frequencies of different category. The numerical tool used for the measurement is the frequency table. bar chart and pie chart are the most common types of graphical tools used to measure the frequencies of different category. Example– How many people are completely satisfied with your company service and how many are not?
- One variable qualitative analysis helps to detect the most frequent category from the given sample. The numerical tool used for the measurement are Mode and the graphical method that is used for such measurements are Bar chart and Pie chart. Example– Which the favourite sport in Europe?
Two variable (Bivariate analysis)
The bivariate analysis helps to measure the association between the two variables. The numerical method used for the measurement are Contingency table and the graphical method that is used for such measurement are the 3D graph of the contingency table (clustered or stacked bar). Example– Does the presence of sulphur in air change according to the presence of other harmful chemicals present in the air?
Mixed variable
Two variable (Bivariate analysis)
Bivariate analysis help to describe the relationship between a binary and a continuous variable. The numerical method used for the measurement are Biserial correlation and the graphical method that is used for such measurement are Box plot. Example- Is the concentration of lead present in the body of humans linked to the humans? sex (M/F)?
Bivariate analysis helps to describe the relationship between a categorical and a continuous variable. The numerical method used for the measurement are univariate description statistics for the qualitative variable within each category of the qualitative variables and the graphical method used is the Box-plot. Example- Does the tails of amphibians differ between five species?
Several Variables
Several variable analysis helps to describe the relationship between one categorical and two quantitative variables.?the graphical method used is Scatterplots (in groups). Example- Does the amount of money spend on this advertisement changes according to the number of views?