As quoted by Mr. Paul Abraham, Executive Vice President, TNS Consumer Intelligence Consulting.
Analytics will always find application in an increasingly "Knowledge-Driven" world. As Information and Communication Technologies capture and bring exponentially growing amounts of data to our desktops, the need for skills to mine this data and move it up the Data-Information-Knowledge-Insight value chain can only grow stronger. The reasons why prospects for analytics will remain bright in the near to medium term is two-fold Given the abundance of trained statistical manpower in India and the relatively low wages, India will remain an attractive destination for the KPO business. Additionally India too will witness the growth in demand for analytics services as user companies (clients) invest in data warehouses to capture and store data, and as markets grow more competitive. They will come under increasing pressure to adopt modern data-supported management practices.
Weblink : Market Research Data Analytics Solutions Provider Bangalore
Tuesday, April 13, 2010
Tuesday, April 6, 2010
Some Analytics Stuff @meltdata
This Section of our blog discusses the important Techniques which are being used in market research industry . These techniques help one in developing a mathematical model to predict various characterstics, behaviours this section will also be useful if we require how to make the right use of statistics, statistics test for market research purposes. We will try to keep it as simple as possible in our approach. We hope that it will be of great use for the researchers and student community.
Of course we will also take the help of spss We will not go into complex arithmetic and derivations behind many of the tests and techniques as it may not be useful.
We assume that the reader to be familiar with basic statistics like mean, mode ,median .Rest other things will be discussed here as and when necessary
One more thing we will use uniformly is 95% confidence interval.
Hypothesis Testing
1.
Generally it may be very confusing to learn various tests and techniques. The golden rule is that whenever you apply any tests you see two things . One is the null hypothesis and the p (sig value) . If the p value is less than 0.05 , your null hypothesis is rejected. So if somebody makes a null hypothesis “ Correlation coefficient is = 0 or there is no correlation between the two variables “ . If the p value is less than 0.05 ,then the null hypothesis is rejected . This implies that there is a correlation between the two variables and your significance however low or high it may be significant”.
Similarly if you have check for regression coefficients . Say our statistical test says the null hypothesis as “All Regression coeff. are zero” and we get a p value > 0,05. That means that “Null hypothesis is true” . This implies that the regression coefficients are not significant. This test of checking all the regression coefficient is called the F test. If we want to check for individual regression coefficnt we need to check their individual p values (which comes from t test of the null hypothesis “Coefficient is zero”. ). In this case also the golden rule holds true.
Let us now use some of the statistical tests. We will systematically study where it is applied . How to do these tests with the spss? How it can be useful in a market research study.
First we will study the tests on Quantitative Data
2 cases exist
1. Normality assumption and homogeneity Assumption are satisfied in the data. Don’t worry we will have a test to check that also!. Normality means it is a bell shape curve while the homogeneity means that the data has a constant variance. Generally these assumption are satisfied in large data by itself.
In these cases we conduct parametric tests.
2. If the above assumptions are not satisfied we would conduct an equivalent Non Parametic test.
Parametric Non-Parametric
1 Sample T-test 1.Sign Test/Wilcoxon Signed Rank test
Paired T-test Sign Test Wilcoxon Signed Rank test
2 Sample T-test 2.Mann Whitney U test/Wilcoxon Sum Rank
ANOVA Kruskal Wallis test
So First of all we do Normality test
1. Plot the graph. See the Histogram . It should look like a Bell curve (Best way)
2. Kolmogorov-Smirnov AND Shapiro-Wilk test will tell you if data is normal. (Nulll hypothesis : Data is Not Normal so p>0.05 means data is normal.). You can do these tests with the spss.
Watch out for more very soon!
Web link : Market Research Data Analytics Solutions provider
Of course we will also take the help of spss We will not go into complex arithmetic and derivations behind many of the tests and techniques as it may not be useful.
We assume that the reader to be familiar with basic statistics like mean, mode ,median .Rest other things will be discussed here as and when necessary
One more thing we will use uniformly is 95% confidence interval.
Hypothesis Testing
1.
Generally it may be very confusing to learn various tests and techniques. The golden rule is that whenever you apply any tests you see two things . One is the null hypothesis and the p (sig value) . If the p value is less than 0.05 , your null hypothesis is rejected. So if somebody makes a null hypothesis “ Correlation coefficient is = 0 or there is no correlation between the two variables “ . If the p value is less than 0.05 ,then the null hypothesis is rejected . This implies that there is a correlation between the two variables and your significance however low or high it may be significant”.
Similarly if you have check for regression coefficients . Say our statistical test says the null hypothesis as “All Regression coeff. are zero” and we get a p value > 0,05. That means that “Null hypothesis is true” . This implies that the regression coefficients are not significant. This test of checking all the regression coefficient is called the F test. If we want to check for individual regression coefficnt we need to check their individual p values (which comes from t test of the null hypothesis “Coefficient is zero”. ). In this case also the golden rule holds true.
Let us now use some of the statistical tests. We will systematically study where it is applied . How to do these tests with the spss? How it can be useful in a market research study.
First we will study the tests on Quantitative Data
2 cases exist
1. Normality assumption and homogeneity Assumption are satisfied in the data. Don’t worry we will have a test to check that also!. Normality means it is a bell shape curve while the homogeneity means that the data has a constant variance. Generally these assumption are satisfied in large data by itself.
In these cases we conduct parametric tests.
2. If the above assumptions are not satisfied we would conduct an equivalent Non Parametic test.
Parametric Non-Parametric
1 Sample T-test 1.Sign Test/Wilcoxon Signed Rank test
Paired T-test Sign Test Wilcoxon Signed Rank test
2 Sample T-test 2.Mann Whitney U test/Wilcoxon Sum Rank
ANOVA Kruskal Wallis test
So First of all we do Normality test
1. Plot the graph. See the Histogram . It should look like a Bell curve (Best way)
2. Kolmogorov-Smirnov AND Shapiro-Wilk test will tell you if data is normal. (Nulll hypothesis : Data is Not Normal so p>0.05 means data is normal.). You can do these tests with the spss.
Watch out for more very soon!
Web link : Market Research Data Analytics Solutions provider
Subscribe to:
Posts (Atom)
Posts
