Mind the Gap.

[passiondrivenstatisticcodeb-blog]

My final Assignment: Correlation of life expectancy and oil consumption versus Income.

by Sebastien Leuba (UK)

Life expectancy and quantities of oil consumed and income relation.

The consumption of oil the last past 10 decades haven’t stop to rise, our income have fallow steadily, like our life expectancy.

Does any real relation exist between is three?

Does Oil Consumption volume and income share a relation?

Does life expectancy increase as oil quantities rise?

Does life expectancy is affected by income?

Data & method

Data

I will use for this hypothesis the Gap Minder project, data are available here data set , the fallowing codebook will help you to understand and read the data.

Methodology

I will use the fallowing data categories:

Incomeperperson 

Lifeexpectancy 

Oilperperson

“Lifeexpectancy” data category has been re-categorized under the fallowing label:

 “50”; “60”;”70”;”80”;”90”

“Oilperperson” data category has been re-categorized by slide of income of “2500” (only for the first) “5000” with an incremental of 5000 after that.

  UNIVARIATE

UNIVARIATE GRAPHIC

BIVARIATE

Correlation

After running a correlation, we can notice  R value close  near '1"( green frame on the above picture)

The P value (red frame) is <0.0001 for 2 of the subset of data

The correlation between oil and life expectancy seen to be weaker than expected, what isn't surprising if you give an other look at the scatter plot above.

Chi Square

A lack of dense data set, make the chi square no usable  for my experiment , but it still returning a good P ratio <0.0001 and definitely in-ply a relation between Life expectancy and  Income.

  Result

Does any real relation exist between is three (oil , Income, life expectancy)?

We can see in our Correlation model , the P value for each is about :<0.0001 ; <0.0001 ; 0.0038.

The R value for each of our examined variable are over 0 (0.60152; 0.54189;0.36267).

We can clearly see a sightly weaken correlation between oil and life expectancy. We will come back on this matter on later study

Does Oil Consumption volume and income share a relation?

We can clear see on the graph (bivariate series)  a clear rise of the oil consumption as the income per person rise. We can see on the correlation table (annotation in orange ) a P value <0.0001 , The R value is 0.54189.

Does life expectancy increase as oil quantities rise

The drawing of a scatter plot show an existing relation but does appear as strong as expected , further the correlation process  show a  "weak but real correlation" between the 2 variables.

A Chi square test show a complete different picture where missing data

A Pvalue 0.4174 , Overall the hypothesis wont be retained by a lack of data.

 Does life expectancy is affected by income?

The drawing of a bivariable graph  show like , the Pvalue of correlation is significant (P=<0.0001) and the R value is 0.6152 .

We can accept the hypothesis that Income affect life expectancy.

Overview 

The hypothesis that oil consumption and higher income will make you live longer is significantly can't be verified, why?

The idea behind oil making you living longer , sound funny some how,(when we know the amount of pollution generated & ecological issues ) but behind is more you get an higher income you can afford more petrol or your country is more industrialised ,  more modern so better public healthcare.

The real problem here is to get a set of data fully populated, sadly the Gap Minder project and the choice of data made din't help. 

The hypothesis will have been certainly more significant if rich data were available at the time of the study but it will be for an other time.

Program Source code

Source Code for Correlation 3 variables

LIBNAME MyDATA "/courses/u_coursera.org1/i_1006328/c_5333" access=readonly; /*collecting data---------------------*/

Data New; SET  MyData.gapminder_pds;

Proc sort ; by COUNTRY; /*Unic identifiant--------------------*/

Proc Freq ; Tables co2emissions Incomeperperson lifeexpectancy oilperperson ; /* Frequency process , tables creation*/

LABEL co2emissions="Co2 Emissions per Tons" /*Label creation easy reading---------*/

Incomeperperson="Income per person in Dollar"

lifeexpectancy="Life Expectancy"

oilperperson="Oil per Person";

/*PROC SORT ; by lifeexepectancy; Moderation on pearson correlation*/

PROC CORR ; VAR lifeexpectancy incomeperperson oilperperson; /*Pearson Correlation*/

run;

Source Code  for chi square.

LIBNAME MyDATA "/courses/u_coursera.org1/i_1006328/c_5333" access=readonly; /*collecting data*/

Data New; SET  MyData.gapminder_pds;

/* Categorization per Age group*/

IF lifeexpectancy LE 50 THEN lifeexpectancy= "50" ; /*Age group less or equal to 50 year old*/

ELSE IF lifeexpectancy LE 60 THEN lifeexpectancy= "60"; /*Age group less or equal to 60 year old*/

ELSE IF lifeexpectancy LE 70 THEN lifeexpectancy= "70"; /*Age group less or equal to 70 year old*/

ELSE IF lifeexpectancy LE 80 THEN lifeexpectancy= "80"; /*Age group less or equal to 80 year old*/

ELSE IF lifeexpectancy LE 90 THEN lifeexpectancy= "90"; /*Age group less or equal to 90 year old*/

/* Categorization Per income  slice*/

IF incomeperperson LE 2500 THEN incomeperperson= "2500" ; /*Income per person group less or equal to 2500$ */

ELSE IF incomeperperson LE 5000 THEN incomeperperson= "5000" ; /*Income per person group less or equal to 5000$ */

ELSE IF incomeperperson LE 10000 THEN incomeperperson= "10000"; /*Income per person group less or equal to 10000$ */

ELSE IF incomeperperson LE 15000 THEN incomeperperson= "15000"; /*Income per person group less or equal to 15000$ */

ELSE IF incomeperperson LE 20000 THEN incomeperperson= "20000"; /*Income per person group less or equal to 20000$ */

ELSE IF incomeperperson LE 25000 THEN incomeperperson= "25000"; /*Income per person group less or equal to 25000$ */

ELSE IF incomeperperson LE 30000 THEN incomeperperson= "30000"; /*Income per person group less or equal to 30000$ */

ELSE IF incomeperperson LE 35000 THEN incomeperperson= "35000"; /*Income per person group less or equal to 35000$ */

ELSE IF incomeperperson LE 40000 THEN incomeperperson= "40000"; /*Income per person group less or equal to 40000$ */

ELSE IF incomeperperson LE 45000 THEN incomeperperson= "45000"; /*Income per person group less or equal to 45000$ */

ELSE IF incomeperperson LE 50000 THEN incomeperperson= "50000"; /*Income per person group less or equal to 50000$ */

ELSE IF incomeperperson LE 55000 THEN incomeperperson= "55000"; /*Income per person group less or equal to 55000$ */

ELSE IF incomeperperson LE 60000 THEN incomeperperson= "60000"; /*Income per person group less or equal to 60000$ */

ELSE IF incomeperperson LE 65000 THEN incomeperperson= "65000"; /*Income per person group less or equal to 65000$ */

ELSE IF incomeperperson LE 70000 THEN incomeperperson= "70000"; /*Income per person group less or equal to 70000$ */

ELSE IF incomeperperson LE 75000 THEN incomeperperson= "75000"; /*Income per person group less or equal to 75000$ */

ELSE IF incomeperperson LE 80000 THEN incomeperperson= "80000"; /*Income per person group less or equal to 80000$ */

ELSE IF incomeperperson LE 85000 THEN incomeperperson= "85000"; /*Income per person group less or equal to 85000$ */

ELSE IF incomeperperson LE 90000 THEN incomeperperson= "90000"; /*Income per person group less or equal to 90000$ */

ELSE IF incomeperperson LE 95000 THEN incomeperperson= "95000"; /*Income per person group less or equal to 95000$ */

ELSE IF incomeperperson LE 100000 THEN incomeperperson= "100000"; /*Income per person group less or equal to 100000$ */

ELSE IF incomeperperson LE 105000 THEN incomeperperson= "105000"; /*Income per person group less or equal to 105000$ */

ELSE IF incomeperperson LE 110000 THEN incomeperperson= "110000"; /*Income per person group less or equal to 110000$ */

Proc sort ; by COUNTRY; /*Unic identifiant--------------------*/

Proc sort ; by lifeexpectancy;    /*third variable ID*/

Proc Freq ; Tables oilperperson*incomeperperson /CHISQ  ; /* X2 declaration + varriable selected from categorical */

LABEL co2emissions="Co2 Emissions " /*Label creation easy reading---------*/

Incomeperperson="Income per person in USD"

lifeexpectancy="Life Expectancy"

oilperperson="Oil per Person/tons/year";

by lifeexpectancy;

  run;

[passiondrivenstatisticcodeb-blog]

Assignment 7 -Chi x2

When what  you are using isnt made for.

I will disregard this Warning  message and I will push to chix2

Code Source for simple chix2:

LIBNAME MyDATA "/courses/u_coursera.org1/i_1006328/c_5333" access=readonly; /*collecting data*/

Data New; SET  MyData.gapminder_pds;

/* Categorization per Age group*/

IF lifeexpectancy LE 50 THEN lifeexpectancy= "50" ; /*Age group less or equal to 50 year old*/

ELSE IF lifeexpectancy LE 60 THEN lifeexpectancy= "60"; /*Age group less or equal to 60 year old*/

ELSE IF lifeexpectancy LE 70 THEN lifeexpectancy= "70"; /*Age group less or equal to 70 year old*/

ELSE IF lifeexpectancy LE 80 THEN lifeexpectancy= "80"; /*Age group less or equal to 80 year old*/

ELSE IF lifeexpectancy LE 90 THEN lifeexpectancy= "90"; /*Age group less or equal to 90 year old*/

/* Categorization Per income  slice*/

IF incomeperperson LE 2500 THEN incomeperperson= "2500" ; /*Income per person group less or equal to 2500$ */

ELSE IF incomeperperson LE 5000 THEN incomeperperson= "5000" ; /*Income per person group less or equal to 5000$ */

ELSE IF incomeperperson LE 10000 THEN incomeperperson= "10000"; /*Income per person group less or equal to 10000$ */

ELSE IF incomeperperson LE 15000 THEN incomeperperson= "15000"; /*Income per person group less or equal to 15000$ */

ELSE IF incomeperperson LE 20000 THEN incomeperperson= "20000"; /*Income per person group less or equal to 20000$ */

ELSE IF incomeperperson LE 25000 THEN incomeperperson= "25000"; /*Income per person group less or equal to 25000$ */

ELSE IF incomeperperson LE 30000 THEN incomeperperson= "30000"; /*Income per person group less or equal to 30000$ */

ELSE IF incomeperperson LE 35000 THEN incomeperperson= "35000"; /*Income per person group less or equal to 35000$ */

ELSE IF incomeperperson LE 40000 THEN incomeperperson= "40000"; /*Income per person group less or equal to 40000$ */

ELSE IF incomeperperson LE 45000 THEN incomeperperson= "45000"; /*Income per person group less or equal to 45000$ */

ELSE IF incomeperperson LE 50000 THEN incomeperperson= "50000"; /*Income per person group less or equal to 50000$ */

ELSE IF incomeperperson LE 55000 THEN incomeperperson= "55000"; /*Income per person group less or equal to 55000$ */

ELSE IF incomeperperson LE 60000 THEN incomeperperson= "60000"; /*Income per person group less or equal to 60000$ */

ELSE IF incomeperperson LE 65000 THEN incomeperperson= "65000"; /*Income per person group less or equal to 65000$ */

ELSE IF incomeperperson LE 70000 THEN incomeperperson= "70000"; /*Income per person group less or equal to 70000$ */

ELSE IF incomeperperson LE 75000 THEN incomeperperson= "75000"; /*Income per person group less or equal to 75000$ */

ELSE IF incomeperperson LE 80000 THEN incomeperperson= "80000"; /*Income per person group less or equal to 80000$ */

ELSE IF incomeperperson LE 85000 THEN incomeperperson= "85000"; /*Income per person group less or equal to 85000$ */

ELSE IF incomeperperson LE 90000 THEN incomeperperson= "90000"; /*Income per person group less or equal to 90000$ */

ELSE IF incomeperperson LE 95000 THEN incomeperperson= "95000"; /*Income per person group less or equal to 95000$ */

ELSE IF incomeperperson LE 100000 THEN incomeperperson= "100000"; /*Income per person group less or equal to 100000$ */

ELSE IF incomeperperson LE 105000 THEN incomeperperson= "105000"; /*Income per person group less or equal to 105000$ */

ELSE IF incomeperperson LE 110000 THEN incomeperperson= "110000"; /*Income per person group less or equal to 110000$ */

Proc sort ; by COUNTRY; /*Unic identifiant*/

Proc Freq ; Tables Incomeperperson*lifeexpectancy /CHISQ  ; /* X2 declaration + varriable selected from categorical */

run;

[passiondrivenstatisticcodeb-blog]

Assignment 6

Anova

When examining the association between quantity of "OILPERPERSON" (quantitative response variable) and life expectancy (categorical variable).

An Analysis of variance (ANOVA) revealed the life expectancy was longer in country where the oil consumption where higher.

For a level of 50 years of life expectancy will be a Mean of 2  tons per person , standard deviation - .

At the opposite  for a level of life expectancy around 90 years the mean consumption of oil 2.383 ton with a Standard deviation 2.514.

F value(2.29)=p 0.0700

[passiondrivenstatisticcodeb-blog]

A quick View

Does a rough view of our data can help to see what we are looking for.

[passiondrivenstatisticcodeb-blog]

Where am I

How far from the end , I could be , to sum the way to :

Assignment 6

Model for presenting results for ANOVA in your blog:

When examining the association between current number of cigarettes smoked (quantitative response variable) and past year nicotine dependence (categorical explanatory variable), an Analysis of Variance (ANOVA) revealed that among young adult smokers (my sample), those with nicotine dependence reported smoking significantly more cigarettes per day (Mean=14.6, s.d. ±9.15) compared to those without nicotine dependence (Mean=11.4, s.d. ±7.43), F(1, 1313)=44.68, p=0001.

Assignment 7

Model for presenting results for Chi-Square tests in your blog:

When examining the association between lifetime major depression (categorical response variable) and past year nicotine dependence (categorical explanatory variable), a chi-square test of independence revealed that among young adults smokers (my sample), those with past year nicotine dependence were more likely to have experienced major depression in their lifetime (36.2%) compared to those without past year nicotine dependence (12.7%), X2 =88.60, 1 df, p=0001.

Assignment 8

Model for presenting results for Correlation in your blog:

Among young adult smokers (my sample), the correlation between number of cigarettes smoked per day (quantitative explanatory variable) and number of nicotine dependence symptoms experienced in the past year (quantitative response variable) was 0.17 (p=.0001), suggesting that only 3% (i.e. 0.17 squared) of the variance in number of current nicotine dependence symptoms can be explained by number of cigarettes smoked per day.

Assignment 9

Evaluate a potential third variable moderator in the context of ANOVA, Chi Square or Correlation and interpret the results. Make sure to comment you code and include your interpretation in a blog post.

IMPORTANT: In order for the automatic grader to process your assignment, you will need to submit your Log as a text file (.txt). Either copy or paste it into a Word Processor, or select Send To:Microsoft Word. Save the resulting document as a text file with the extension .txt, and Submit it via the Assignment Page.

----------------------------------------------------------------------------------------

[passiondrivenstatisticcodeb-blog]

Assignment 5

I used the assignment 5 to put at test directly my hypothesis and to get a visual of the relation via scatter plot , the fallowing scatter show a poor relation between Oil consumption and life expectancy.

Futher investigation will be needed to develop the current hypothesis

[passiondrivenstatisticcodeb-blog]

Assignment 4

The Assignment 4 with an  income distribution.

[passiondrivenstatisticcodeb-blog]

That is what I'm looking for.

This Question has been revisited , you better look at latest post

    I selected a certain amount of data from the "Gapminder" ,I going to work on the fallowing questions:

Does life expectancy is influenced by the quantity of oil consumed and the quantity of CO2 emitted? 

Relation  between income, life expectancy and oil consumption ,?

[passiondrivenstatisticcodeb-blog]

What am I looking at?

[passiondrivenstatisticcodeb-blog]

Data Import , painful maybe not...

When everything fail, nothing better than a old fashion .CSV file

After multiple attempt  with "LIBNAME My data" and a log full of error and spending over hours reading post on coursera web site nothing was coming out of it.

I decided to import manually my data with an old fashioned CSV (commas separated value) file.

The Importation run pretty smoothly much like I was doing a Excel Import file, always a step a the time!

[passiondrivenstatisticcodeb-blog]

Test Post

That will the real kick start of "Passion driven statistic" Compulsory Blog associated with the course.

As usual I'm already late for flooding the coursera forum  of trouble questions of mine.