---
title: "Working with Network Survey Data in R"
author: "Marcel Mallow"
format: html
---

## Getting our collected survey data into R

```{r}
  # if you are pretty sure you did not install any of the below packages to your R / RStudio environment before, then please run the install-commands
 # otherwise just go to the next code block

#install.packages("tidyverse")
 # install.packages("dplyr")
 # install.packages("tidyselect")
 # install.packages("stringr")  


 # otherwise load / import the packages you will need for the data import and cleaning / wrangling in this session
 

library (tidyverse)
library (tidyselect)
library (stringr)
library (dplyr)


```


```{r}
    # otherwise load / import the packages you will need for the data import and cleaning / wrangling in this session

    getwd() #tell you what your working dir is

    # setwd() #lets you set a new working directory, to the folder where your .csv resides
```
When you are loading the network survey .csv make sure it is in your current R working directory, otherwise you will have to specify the path.
```{r}
# set this to your working directory
# setwd("~/surveyDataImport/waves")

# responsesFile <- "sampleNetworkSurvey.csv" #set this to the filename of your downloaded Qualtrics .csv file 


responsesFile <- "https://raw.github.com/MarMall/NetworkApproaches/main/Instructions/Labs/data/sampleNetworkSurvey.csv" #set this to the filename of your downloaded Qualtrics .csv file 

```


```{r}

# #remove row(s) with question
# responses <- responses [-c (1), ] 
# head(responses)

#better way to avoid row with question recoding everything into character variables
#so we don't have to change them all back again later

#setwd("~/surveyDataImport/waves")

# readr::

responses <- read.csv (responsesFile, header = TRUE, nrows = 1) #only the first row
variables <- colnames (responses) # then set this to our variable names

#variables
#head(responses)

# variables[1] <- "ID"
responses <- read.csv (responsesFile, header = FALSE, skip = 2)
colnames (responses) <- variables

## Here you have your dataframe of all the responses
# have a look inside
responses 
```
## Isolating our network relations of interest

```{r QID149-Friendship}

## Now lets recode the first question results into network format

#responses to first question -- selection by question number

## exclude 99 as this is "Nobody"
QID149 <- responses %>%
  select (c (
            #"ID",
             #"recipientLastName",
             # "QID169_1",
             vars_select (names (responses),
                          starts_with ("QID149", ignore.case = TRUE),
                          -starts_with ("QID149_99", ignore.case = TRUE) #filter out NOBODY
                          )))


# QID149 <- QID149[2:7]



QID149 <- QID149[1:14] #we only have 14 responses, so we work only with those.
#QID149
```

```{r QID128-Influence(Top10-Influencers)}
## exclude 99 as this is "Nobody"
QID128 <- responses %>%
  select (c (
            #"ID",
             #"recipientLastName",
             # "QID169_1",
             vars_select (names (responses),
                          starts_with ("QID128", ignore.case = TRUE),
                          -starts_with ("QID128_99", ignore.case = TRUE)
                          )))


# QID128 <- QID128[2:7]

QID128 <- QID128[1:14] # again, we only had 14 responses to the survey
#QID128


QID128 <- QID128 %>%
  # Recoding NAs to 0  
  mutate (across (vars_select (names (QID128),
                               contains ("QID128", ignore.case = TRUE)),
          ~ replace (., is.na (.), 0))) %>%
  ## AND making the data binary
    mutate (across (vars_select (names (QID128),
                               contains ("QID128", ignore.case = TRUE)),
          ~ replace (., . != 0, 1))) #this ensures the data is binary


dim(QID128) # nice, this is squared 14x14

#QID128
```

Now we have to turn it all into matrix format, but that is easy with R...

```{r turningIntoMatrixFormat}
# 
# QID149 # friendship relation
# QID128 # influence

QID149_mat <-  as.matrix(QID149)

#QID149_mat
is.matrix(QID149_mat)

# checking the dimensions of the matrix
#dim(QID149_mat) # should be symmetric

##same for the influence relation

QID128_mat <-  as.matrix(QID128)

# # checking the dimensions of the matrix
# dim(QID128) # should be symmetric

#QID128_mat
is.matrix(QID128_mat)

```


OK, we now have two network relations in matrix format, and you can do the procedure for as many network relations as you like.
However, this might already be a good moment to already save your matrices in R, so that you can access them in later R-Sessions.

## Saving Network relation matrices

```{r SavingNetworkRelation}

#needs to be in .RData format

# will save directly into your workspace / working directory (getwd())
save (QID149_mat,
        file = paste0 ("Friendship_q149"  ,".Rdata"))

save (QID128_mat,
        file = paste0 ("Influence_q128"  ,".Rdata"))


```