Objective

Script that I use to check and predict the cost and results of Facebook Campaigns. The campaigns are based upon “click in external links”, so the results that I expect are how much do I need to invest in order to get some numbers of clicks.

### Load data
df = read.csv("./data/cmkbr_full.csv", header = TRUE, sep = ",")

Pre processing

Exploratory analysis

  1. Summary
##    Resultados        Alcance      Custo.por.resultados Valor.gasto..BRL.
##  Min.   :  11.0   Min.   :  278   Min.   :  0.1665     Min.   :  7.71   
##  1st Qu.:  61.0   1st Qu.: 2852   1st Qu.:  0.2834     1st Qu.: 27.48   
##  Median : 142.0   Median : 5848   Median :  0.3704     Median : 49.61   
##  Mean   : 340.1   Mean   :11719   Mean   :  5.2327     Mean   :118.85   
##  3rd Qu.: 368.5   3rd Qu.:14021   3rd Qu.:  0.5400     3rd Qu.:100.00   
##  Max.   :2342.0   Max.   :64245   Max.   :263.0000     Max.   :932.56   
##  Pessoas.executando.uma.ação
##  Min.   :  15.0             
##  1st Qu.:  79.0             
##  Median : 183.0             
##  Mean   : 454.7             
##  3rd Qu.: 624.0             
##  Max.   :2040.0
  1. Plot of distribution

  1. Correlation Summary
##                              Resultados      Alcance Custo.por.resultados
## Resultados                  1.000000000 0.9295720543         0.0084991563
## Alcance                     0.929572054 1.0000000000         0.0001948107
## Custo.por.resultados        0.008499156 0.0001948107         1.0000000000
## Valor.gasto..BRL.           0.755092352 0.8148605265        -0.0112340984
## Pessoas.executando.uma.ação 0.839330668 0.8226345495        -0.0201093177
##                             Valor.gasto..BRL. Pessoas.executando.uma.ação
## Resultados                          0.7550924                  0.83933067
## Alcance                             0.8148605                  0.82263455
## Custo.por.resultados               -0.0112341                 -0.02010932
## Valor.gasto..BRL.                   1.0000000                  0.65861092
## Pessoas.executando.uma.ação         0.6586109                  1.00000000
  1. Plot two Comparisons

Remove or not the Outliers

Regression Model

Residuals analysis

## 
## Call:
## lm(formula = df$Resultados ~ df$Valor.gasto..BRL.)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -948.42  -51.50  -17.46   17.60  836.81 
## 
## Coefficients:
##                      Estimate Std. Error t value Pr(>|t|)    
## (Intercept)             5.797     35.157   0.165     0.87    
## df$Valor.gasto..BRL.    3.192      0.191  16.708   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 211.3 on 51 degrees of freedom
## Multiple R-squared:  0.8455, Adjusted R-squared:  0.8425 
## F-statistic: 279.2 on 1 and 51 DF,  p-value: < 2.2e-16