Tuesday, April 13, 2010

Scope of Data Analytics In the World

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 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
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

Thursday, March 18, 2010

Union Budget 2010-11: How it's likely to cause price rise, petro and gold prices would impact inflation

The Union finance minister, Pranab Mukherjee’s budgetary proposals for the year 2010-11 may help the economy to achieve the magic growth figure which is targeted at 8.5%, but little it can do to vanish the price inflation worries. In fact certain proposals in the Budget would lead to further rise in prices of essential commodities.

The proposal to restore the basic customs duty of 5% on crude petroleum, 7.5% on diesel and petrol and 10% on other refined products was justified by finance minister by saying that the increse in custom duty is the roll back to the level which was reduced in June 2008 as the prices of crude was pretty high about $120 / barrel at that time but tnow the prices are softened. And moreover it was the time to move on the path of fiscal consolidation.

The global prices of crude oil had softened in the wake of the global financial crisis, but now it has shown a rising trend. Currently the global prices are hovering around US$ 80 per bbl. As India depends upon crude oil imports to the extent of 70%, these measures of the government would definitely translate into further price inflationary pressure on the economy. Following the announcement of the budgetary proposals, the Indian oil companies have decided to upscale the prices. Diesel prices in Delhi is likely to increase by Rs 2.55 and that of petrol by Rs 2.71 a litre. Thus not only public transport would become costlier, but also the transportation of essentialcommodities. Already the country is reeling under the impact of soaring prices.

Another folly in the budgetary proposal is to increase the customs duty on gold and platinum from Rs 200 per 10 grams to Rs 300 per 10 grams and on silver from Rs 1,000 per kg to Rs 1,500 per kg. This caused the rise gold prices and the spillover effect was on the prices of other commodities as well. Gold has now become an important area for investment and in India, particularly the attachment to this precious metal is due to the cultural ethos.

Instead of raising the custom duties the FM should rollover some of the incentives which was given to the manufacturing sector at  the time of financial crisis and as in current scenario they have registered a growth of around 8.9%, some of these incentives could have been withdrawn in a calibrated manner.

But I must admit that this a strong step taken by the finance minister without worrying about the reaction of opposition parties. Definitely in long term it will be beneficial for shooting up the growth rate of GDP of country. But short term effect will be double digit inflation in few months.

Weblink : Market Research Data Analytics Solutions Provider Bangalore

Saturday, March 13, 2010

IS excess saving is a virtue or a vice ?

Hi Readers

Here I am trying to explorethe implications of excess saving. Generally our perceptions are like individual saving will result in better standard of living in future, means thriftness will help in improve saving thus better life in future. Tha anmoly arises because of paradox of saving/ thriftness as I will explain in detail in subsequent lines.

As we know by IS / goods equilbrium

Total output  ( Y ) = Expected Demand ( Z )..................Goods market equilbrium

Thus  Y = C + I + G (In case of close economy)

Y = [ Co + c (Y - T)  ] + G + I

Total saving = Private saving + Public saving

Private saving   S = ( Y - C - T)  = G + I + T.............and we know all the 3 factors are independent of thriftness. So overall private saving does not change.

But Y = [ Co + c (Y - T) ] + G + I  decreased.

Classical Model

In classical model we assume Y is constant as in long run, economy is producing at full employment level .

So Total Saving  S = Y - C - G and here due to thriftness C decreased but Y is constant, thus overall saving decreaeses. This is paradox of Saving.

So thriftness can lead to decrease or unchanged Saving.

Weblink : Market Research Data Analytics Solutions Provider Bangalore

Wednesday, March 10, 2010

Why some countires are so rich while others are mired in poverty??

 Before starting to explain the phenomenon as to why their a a difference in status of countries. Firstly we need to understand in what terms we are saying that a country is rich and other is poor. Is it standard of living, employment, aggregate income of a country or inflation and several other economic indicators.

To clear the point it is standard of living of people of country which is stated in terms of real GDP per capita i.e. on an average income of a person in a country.

Before going into details, we need to understand a comprehensive model named as Solow Growth Model named after the famous economist Robert Solow. We need to clearly understand the assumptions prior to diving into the mode.

  • Production function exhibits Diminishing Marginal return to Capital
  • We assume 2 inputs in the production function - Capital and Labor
  • Two other important inputs are Natural Resources and Human capitsl (but at this point of time we will not consider these two inputs)
  • We will assume that growth in "Technology"  is  not there.
Production function :  Y =  A F(K,L)

Assuming y = Y/L and k = K/L, we can rewrite the equation as y = f (k) as we are interested in analyzing output per labor.

Also as we know that  Saving = Investment = s f (k) where s = saving rate

The change in capital stock per worker :

dk = Investment (sf(k)) - depreciation in capital (d*k) - rate of increase in population growth (n) * k

    Increase in population requires additional capital
    We need to maintain the condition on existing machines/ capital stock
    Investment on capital will help in increasing capital stock 

At steady state dk = 0
hence,   sf (k) =  ( n + d ) k

Key findings :

1. At steady state output / worker and capital / worker is constant i.e. growth is zero. 

2. You can see clearly from the graph that if saving rate is high :  Steady state capital / worker and output / worker will be high.

3. While if rate of depreciation of capital or population growth (n) is high standard of living will be low. 

4. Steady state growth in total output of a country is n i.e. population growth

So countries with high population growth and low saving rate are poorer

      ######  Keep in mind that we excluded  the effect of technological growth in this model. ######

Weblink : Market Research Data Analytics Solutions Provider Bangalore

Tuesday, March 9, 2010

Mathematical Model for Deriving Equilbrium Income and Interest Rate

In this article we are going to explain the mathematics behind deriving Aggregate Demand from IS - LM Model (where goods and capital market are both in equilbrium simultaneously).

The intersection of IS and LM schedules determines the equilbrium income and the equilbrium interest rate. We will use the equationg derived for IS and LM schedules

IS schedule : Y = α  ( Â – b i )

LM schedule :  i = (1/h ) ( kY - M / P )

Solving the two schedules we will derive the equilbrium income

Equilbrium Income       Y = γ Â + γ ( b / h) ( M / P)

We can calculate equilbrium interest rate by substituting  eqilbrium income in Lm schedule.

Weblink : Market Research Data Analytics Solutions Provider Bangalore

Friday, March 5, 2010

Equilbrium in the Goods Market and Money Market

IS - LM macroeconomic model helps us to develop a sound understanding of how equilbrium is approached in our economy. Here we will understand how equilbrium is acheived in a closed economy.

A macroeconomic model that graphically represents two intersecting curves, called the IS and LM curves. The investment/saving (IS) curve is a variation of the income-expenditure model incorporating market interest rates (demand for this model), while the liquidity preference/money supply equilibrium (LM) curve represents the amount of money available for investing (supply for this model).

The interest rate and the level of output are determined by the interaction of money (LM) and goods (IS) markets. But we should remember that we have adopted few assumptions while developing IS - LM model.

Key Assumptions:

  • Price level is assumed to be constant

  • Firms are willing to supply whatever amount of output is demanded at that price level.

  • We assumed short rum aggregate supply curve to be flat

We will explain the mathematics behind the model in the subsequent blog.

Weblink : Market Research Data Analytics Solutions Provider Bangalore