Is data the new oil?

Once upon a time, the oil used to rule the global economy. It was an immensely precious asset of the 18th century. Without oil, economies used to shrink, governments used to tumble, and progress used to halt. Welcome to the future – an innovative particular wonder is amazing the world, the advent of technology and the flow of data marked the beginning of the new global dominator. Around a decade ago, we heard what Clive Humby said – “Data is the new oil.” According to Clive, data is a commodity and if treated just like oil, data is also a resource of no value if it remains unrefined.

Comparing data with oil

  • For decades oil has fueled flourishing of the industrial economy, and now data is empowering information economy.
  • Previously the economic wealth was dependent upon how much natural resources one possesses, but today, wealth in modern economies is directly proportional to the amount of data they have.
  • If the oil is a resource used to extract energy, data is used to extract information.
  • There was a time when oil led to the transportation revolution and now data is revolutionizing the technology like driverless cars, hyperloop, AI and machine learning.
  • Oil is a natural resource hence is finite, on the contrary, data is infinite and will keep on growing.
  • The flow of oil and data is similar in many instances including drilling mechanisms. We get oil after ferocious drilling, similarly, data needs to be drilled to extract value.
  • We get plastics, petrochemicals, lubricants, gasoline, and many such byproducts from oil. Whereas, data gives us algorithms, valuable insights, and innovative options for technology development.
  • Oil is a rival good as it cannot be used twice. Conversely, data being a non-rival good can be used as many times as you want.
  • The lifecycle of oil is determined by the process used to extract, refine, and distribute it. Whereas, the lifecycle of data is defined by its relation with itself and with the other data through feedback loops.

Data refinement

Similar to oil, data needs refinement too

Do you think economies having plenty of oil resources are the wealthiest? Sadly, no… but the countries having advanced potential to refine the oil are surely amongst the wealthiest economies. Just like oil, the value of data increases in many folds when it is refined. A refined data gives us information, this information is then converted to knowledge and the knowledge thus gained is utilized to make decisions for getting expected results.

Data drilling is not something related to excel analytics but it is beyond that, it covers 5 V’s – Volume, Velocity, Variety, Veracity, and Value. In addition to that, data drilling adds 3 A’s value to the raw data. A refined data in the form of ‘Analytics’ is smartly transformed into sequences of instructions resembling ‘Algorithm’, these sequences of instructions are further developed into a consistent platform for running multiple programs like ‘Applications’.

If data is the new oil, then data analysis is that combustion engine which generates mechanical power to derive insights and gain knowledge. Gone are the days when descriptive and diagnostic analytics used to rule the data analysis regime. It’s time we move on to predictive, perspective, and cognitive analysis.

Data analysis

Be it any data, the basic methods used for data analysis are statistical analysis. Majority of quality management methods as if Six Sigma happen to be statistic intensive applying a variety of statistical techniques for data analysis and confirm the extent of deviation from the standard mean. These analytical techniques are very much useful in the long run for uncovering the precise information that will lead to decisions making and effective results.

Use correlation analysis to drill survey data

Definition: Correlation analysis is defined as a statistical approach used to determine the relationship between the quantitative variables or categorical variables. With the help of correlation analysis, we can prove the relationship between two continuous variables, in short, it is a study conducted to understand how two variables are correlated.

Example: two variables like your body weight and daily calorie intake have a higher correlation. On the contrary, dog names and the brand of dog biscuits preferred by dogs will have a very low correlation.

When you want to determine correlation through data analysis, there are two types of data you need to work with

Univariate Data: When you want to work with a single variable, you need to measure the central tendency for inquiring about the representative data and know the deviation around the central tendency via dispersion, measure the shape and size of the distribution through skewness and measure the concentration of data at the central position through kurtosis. Thus, data relating to a single variable is called Univariate data.

Bivariate Data: Correlation analysis is more about studying the relationship between two variables at the same time. For instance, the price of a product and average sales of the product, or age and blood pressure of a person. Thus, two characters of the same entity when measured simultaneously with the help of statistical analysis then we term it as a bivariate data.

If there is any kind of correlation between two variables, then whenever there is a systemic change is one variable, the other variable also changes. Thus, the variables alter together over a period of time. The correlation thus found can be either positive or negative depending on the measured numerical values.

  • A positive correlation exists when due to an increase in any one of the variables, the other variable also starts increasing ensuring positive correlation between them. Example: body weight of male and female are positively correlated
  • A negative correlation exists when due to the increase in any one of the variables, the other variable starts decreasing ensuring a negative correlation between them. Example: an increase in price leading to a decrease in sales.

Four methods to analyze correlation results

In statistical methods, the correlation coefficient “r” measures the strength, direction, and extent of the relationship between two variables, where the value of “r” will always hinge between +1 and -1. Remember, it’s futile to calculate the correlation if there is no relation between the two variables as correlation only applies to linear relationships. Conversely, if there is a strong relationship between the two variables but it is not linear then the correlation received may be misleading. Therefore, it is advisable that before conducting correlation research by using any of the correlation coefficient methods always examine the scatter plot first. Here are some of the commonly used coefficient correlation methods.

  1. Scatter Diagram Method

Scatter diagram method is a naive approach in correlation analysis used to find the correlation between two variables. Relationship between the two variables is diagrammatically presented to understand how closely they are related to each other.

    • Also, called a scatter plot, scatter graph, or correlation chart
    • The diagram or the chart has two variables along its ‘x’ and ‘y’ axis out of which one is independent and the other is the dependent variable
    • It’s easy to predict the behavior of the de independent variable depending on the measure of the independent variable,
    • According to the type of correlation scatter diagrams are divided into Scatter diagram with no correlation, scatter diagram with moderate correlation and scatter diagram with a strong correlation.

Scatter-Diagram

  1. Pearson Correlation Coefficient

Also called as the ‘product-moment correlation coefficient (PMCC) or simply ‘correlation’. It is defined as a number between -1 and 1 indicating the extent to which the two variables are linearly related.

    • Pearson correlation method is suitable for metric variables which also includes dichotomous variables.
    • The correlation is always denoted as ‘“r
    • r’ is independent of the unit of measurement. An example is one variable is in inches and the other is in quintal then also the value of Pearson’s coefficient correlation does not change.
    • Pearson correlation formula

ATTACHMENT DETAILS Pearsons-correlation-coefficient

  1. Spearman Rho Coefficient

It is a non-parametric version of the Perasnos’s coefficient correlation. This method is utilized to measure the strength and direction of relationship or association existing between the two variables.

    • Greek letter rho (ρ) is used to denote the Spearman’s Correlation Coefficient
    • It is also denoted by the symbol rs
    • Spearman Rho is used either for ordinal variables or for continuous data that has failed the assumptions necessary for conducting the Pearson’s correlation coefficient.
    • Spearman Rho Formula

Spearman-Rho

  1. Least Squares Method

It is a mathematical problem used to find the degree of correlation between the two variables by using the square root of the product of two regression coefficient that of ‘x’ on ‘y’ and ‘y’ on ‘x’.

    • The least square method ensures to make the total of the square of the errors as low as possible
    • Also called the Line of Best Fit
    • It is used to calculate the mean of ‘x’ and ‘y’ values.
    • The Least Squares Method formula

Least_Square_Method

How to use correlation analysis in QuestionPro surveys

  • Login into your QuestionPro account
  • Go to ‘Surveys’

Data-Analysis

  • Find the survey you want to analyze and open it
  • Go to ‘Analytics’ press the button

correlation-analysis-in-surveys

  • Go to Analytics > ‘Correlation Analysis

  • The Highest positive correlation between employees and the time they are working with the company (+0.79)
  • The Lowest negative correlation is between employees and their level of satisfaction with the company (-0.62)

Data is like oil. Similar to oil, people are finding new ways and sources to generate data. With evolving technology, we are finding new and better ways to collect data efficiently and options to refine this huge amount of data with specialized techniques.

Do you know a barrel of crude oil costs less compared to a barrel of jet fuel? This is because refining crude oil yields jet fuel which further increases in value by 33% to 35%. Data is the same, you need to process it to gain maximum value. We, at QuestionPro, make it easier for you to drill and refine your survey data.