Difference between revisions of "An initial path towards statistical analysis"

From Sustainability Methods
Line 181: Line 181:
 
== Network analysis ==
 
== Network analysis ==
 
<imagemap>Image:Statistics Flowchart - Network Analysis.png|center|800px|
 
<imagemap>Image:Statistics Flowchart - Network Analysis.png|center|800px|
poly 328 368 -4 688 332 1004 636 696 624 680 [[Bipartite|Big problems for later|]]
+
poly 336 364 4 684 340 1000 644 692 632 676 [[Bipartite|Big problems for later|]]
 
poly 1064 372 732 696 1060 1004 1372 688 [[Tripartite|Big problems for later]]
 
poly 1064 372 732 696 1060 1004 1372 688 [[Tripartite|Big problems for later]]
 
</imagemap>
 
</imagemap>

Revision as of 08:36, 10 March 2021


Start here with your data! This is your first question.

Error: Image is invalid or non-existent.



Data inspection example

Do you have several continuous variables without clear dependencies? (?)
Yes, I have several continuous variables without clear dependencies!
No, my variables have clear dependencies!
R commands: str, summary, head, tail
Example: Inspecting the swiss dataset



Univariate statistics

My variables have clear dependencies!


Does you data contain at least one categorical variable?
Yes, my data contains at least one categorical variable?
No, my data contains only continuous variables?

No, my variables have clear dependencies!


R commands: str, summary, head, tail, ordered(dataset$variablename, c(levels = “level1”, level2”...)) 

Relevant figures: hist(), boxplot()

 
Yes, my data contains at least one categorical variable?


Categorical variables

Does you data contain only categorical variables?
Yes, my data contain only categorical variables?
No, my data does contains continuous and categorical variables?

Yes, my data contain only categorical variables?


Chi-Square test

R commands: table, chisq.test( x, correct=TRUE)
Relevant figures: (stacked) bar charts : barplot(), pie(), bipartite

No, my data does contains continuous and categorical variables?


Categorical and continuous data

R commands: quantile(), str, summary,  

Does you categorical dependent variables have 1-2 factor levels?
Yes, my categorical dependent variables has 1-2 factor levels!
No, my categorical dependent variables more than 2 factor levels!

Yes, my categorical dependent variables has 1-2 factor levels!


t-test

R commands: t.test, t_test Relevant figures:

No, my categorical dependent variables more than 2 factor levels!


Analysis of Variance

R commands: aov, Anova, ezAnova, var.test(), lm
Relevant figures: boxplot()

Is your dependent variable normally distributed?

R commands: ks.test, shapiro.test, hist 

Yes, my dependent variable is normally distributed!
No, my dependent variable is binomial distributed!
No, my dependent variable is Poisson distributed!

Yes, my dependent variable is normally distributed!


Gaussian Anova

R commands: aov, lm
Relevant figures: boxplot


No, my dependent variable is not normally distributed!


Is your dependent variable binomial or Poisson

No, my dependent variable is Poisson distributed!


Dependent variable is count data

R commands: glm,
Relevant figures: plot

No, my dependent variable is binomial distributed!


Dependent variable is 0/1 or proportions

R commands:
Relevant figures:



Type III Anova

R commands: Anova(car)
Relevant figures: boxplot

Dependent variable not normally distributed

Dependent variable is count data

R commands: glm
Relevant figures: plot

Dependent variable is 0/1 or proportions

R commands: glm
Relevant figures:

Are there random factor variables?

Random factors

R commands: glmer, glmmPQL
Relevant figures:


No, my data contains only continuous variables?


Continuous variables

Non dependent relations?

Correlations

Clear dependent relations

Regression models

Dependent variable normally distributed

Linear Regression

Dependent variable not normally distributed

Generalised linear model

Dependent variable is count data

Dependent variable is 0/1 or proportions

R commands:
Relevant figures:

Yes, I have several continuous variables without clear dependencies!


Multivariate statistics

Statistics Flowchart - Clustering, Networks, Ordination.png
About this image


R commands:
Relevant figures:

Data is classified into groups

R commands:
Relevant figures:

Cluster analysis

DIFFERENCE BETWEEN SUPERVISED AND UNSUPERVISED?? DISTINCTION NOT MADE IN THE CLUSTERING ENTRY

Supervised Classification Unsupervised ClassificationStatistics Flowchart - Cluster Analysis.png
About this image

R commands:
Relevant figures:


Supervised classification

R commands:
Relevant figures:

Unsupervised classification

R commands:
Relevant figures:

Network analysis

Big problems for later| Big problems for laterStatistics Flowchart - Network Analysis.png
About this image

R commands:
Relevant figures:

Bipartite

R commands: is_bipartite(graph)


Relevant figures: make_bipartite_graph(types, edges, directed = FALSE)

Tripartite

R commands:
Relevant figures:

Ordinations

Euclidean distances Jaccard distancesStatistics Flowchart - Ordination.png
About this image

R commands:
Relevant figures:

Linear based ordinations

R commands:
Relevant figures:

Correspondance analysis

R commands:
Relevant figures:



More than 2 categorical variables


Is your dependent variable normally distributed?
Is your dependent variable not normally distributed?

My data consists only of categorical variables

Does your independent variable contain only 1 or 2 groups?
Does your independent variable contain more than 2 groups?

Does your independent variable contain more than 2 groups?


Is your dependent variable normally distributed?
Is your dependent variable not normally distributed?



Does your independent variable contain more only 1 or 2 groups?



My data consists only of categorical variables


Multivariate statistics



Resterampe

[[At least one continuous and one categorical variable| More than 2 groups Analysis of Variance
Dependent variable normally distributed
INSERT TYPE II
INSERT RANDOM FACTOR
INSERT LMM

Dependent variable not normally distributed

Type III Anova

Dependent variable is count data

Dependent variable is 0/1 or proportions

Random factors