Walmart

Importar base de datos 

#file.choose()

stores <- read.csv("/Users/ErickaMtz/Downloads/stores (1).csv")
features <- read.csv("/Users/ErickaMtz/Downloads/features (1).csv")
train <- read.csv("/Users/ErickaMtz/Downloads/train (1).csv")
test <- read.csv("/Users/ErickaMtz/Downloads/test (1).csv")

Instalar paquetes y llamar librerías 

#install.packages("dplyr")
library(dplyr)
## 
## Attaching package: 'dplyr'
## The following objects are masked from 'package:stats':
## 
##     filter, lag
## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

Entender la base de datos 

summary(stores)
##      Store        Type                Size       
##  Min.   : 1   Length:45          Min.   : 34875  
##  1st Qu.:12   Class :character   1st Qu.: 70713  
##  Median :23   Mode  :character   Median :126512  
##  Mean   :23                      Mean   :130288  
##  3rd Qu.:34                      3rd Qu.:202307  
##  Max.   :45                      Max.   :219622
count(stores, Type,  sort = TRUE)
##   Type  n
## 1    A 22
## 2    B 17
## 3    C  6
str(stores)
## 'data.frame':    45 obs. of  3 variables:
##  $ Store: int  1 2 3 4 5 6 7 8 9 10 ...
##  $ Type : chr  "A" "A" "B" "A" ...
##  $ Size : int  151315 202307 37392 205863 34875 202505 70713 155078 125833 126512 ...
summary(features)
##      Store        Date            Temperature       Fuel_Price   
##  Min.   : 1   Length:8190        Min.   : -7.29   Min.   :2.472  
##  1st Qu.:12   Class :character   1st Qu.: 45.90   1st Qu.:3.041  
##  Median :23   Mode  :character   Median : 60.71   Median :3.513  
##  Mean   :23                      Mean   : 59.36   Mean   :3.406  
##  3rd Qu.:34                      3rd Qu.: 73.88   3rd Qu.:3.743  
##  Max.   :45                      Max.   :101.95   Max.   :4.468  
##                                                                  
##    MarkDown1        MarkDown2           MarkDown3           MarkDown4       
##  Min.   : -2781   Min.   :  -265.76   Min.   :  -179.26   Min.   :    0.22  
##  1st Qu.:  1578   1st Qu.:    68.88   1st Qu.:     6.60   1st Qu.:  304.69  
##  Median :  4744   Median :   364.57   Median :    36.26   Median : 1176.42  
##  Mean   :  7032   Mean   :  3384.18   Mean   :  1760.10   Mean   : 3292.94  
##  3rd Qu.:  8923   3rd Qu.:  2153.35   3rd Qu.:   163.15   3rd Qu.: 3310.01  
##  Max.   :103185   Max.   :104519.54   Max.   :149483.31   Max.   :67474.85  
##  NA's   :4158     NA's   :5269        NA's   :4577        NA's   :4726      
##    MarkDown5             CPI         Unemployment    IsHoliday      
##  Min.   :  -185.2   Min.   :126.1   Min.   : 3.684   Mode :logical  
##  1st Qu.:  1440.8   1st Qu.:132.4   1st Qu.: 6.634   FALSE:7605     
##  Median :  2727.1   Median :182.8   Median : 7.806   TRUE :585      
##  Mean   :  4132.2   Mean   :172.5   Mean   : 7.827                  
##  3rd Qu.:  4832.6   3rd Qu.:213.9   3rd Qu.: 8.567                  
##  Max.   :771448.1   Max.   :229.0   Max.   :14.313                  
##  NA's   :4140       NA's   :585     NA's   :585
str(features)
## 'data.frame':    8190 obs. of  12 variables:
##  $ Store       : int  1 1 1 1 1 1 1 1 1 1 ...
##  $ Date        : chr  "05/02/2010" "12/02/2010" "19/02/2010" "26/02/2010" ...
##  $ Temperature : num  42.3 38.5 39.9 46.6 46.5 ...
##  $ Fuel_Price  : num  2.57 2.55 2.51 2.56 2.62 ...
##  $ MarkDown1   : num  NA NA NA NA NA NA NA NA NA NA ...
##  $ MarkDown2   : num  NA NA NA NA NA NA NA NA NA NA ...
##  $ MarkDown3   : num  NA NA NA NA NA NA NA NA NA NA ...
##  $ MarkDown4   : num  NA NA NA NA NA NA NA NA NA NA ...
##  $ MarkDown5   : num  NA NA NA NA NA NA NA NA NA NA ...
##  $ CPI         : num  211 211 211 211 211 ...
##  $ Unemployment: num  8.11 8.11 8.11 8.11 8.11 ...
##  $ IsHoliday   : logi  FALSE TRUE FALSE FALSE FALSE FALSE ...
summary(test)
##      Store            Dept           Date           IsHoliday      
##  Min.   : 1.00   Min.   : 1.00   Length:115064      Mode :logical  
##  1st Qu.:11.00   1st Qu.:18.00   Class :character   FALSE:106136   
##  Median :22.00   Median :37.00   Mode  :character   TRUE :8928     
##  Mean   :22.24   Mean   :44.34                                     
##  3rd Qu.:33.00   3rd Qu.:74.00                                     
##  Max.   :45.00   Max.   :99.00
str(test)
## 'data.frame':    115064 obs. of  4 variables:
##  $ Store    : int  1 1 1 1 1 1 1 1 1 1 ...
##  $ Dept     : int  1 1 1 1 1 1 1 1 1 1 ...
##  $ Date     : chr  "02/11/2012" "09/11/2012" "16/11/2012" "23/11/2012" ...
##  $ IsHoliday: logi  FALSE FALSE FALSE TRUE FALSE FALSE ...

Observaciones

  1. “features”, “test” y “train” tiene la fecha como caracter
  2. “features”, hay NA’s en más de la mitad de los registros de Markdown (del 1 al 5).
  3. En “features”, hay 585 NA’s en CPI y Unemployment, y hay 585 registros de IsHoliday = TRUE ¿Tiene relación? R: Sin relación.
  4. “train” tiene ventas semanales negativas

Herramienta “El Generador de Valor de Datos”

Paso 1. Definir el área del negocio que buscamos impactar o mejorar su KPI.
El departamento de mercadotecnia de EUA (con muestra de 45 tiendas) en el indicador de ventas semanales
Paso 2. Seleccionar plantilla (-s) para crear valor a partir de los datos de los clientes
Visión / Segmentación / Personalización / Contextualización
Paso 3. Generar ideas o conceptos específicos
Elaborar un modelo predictivo de ventas semanales
Paso 4. Reunir los datos requeridos
Elaborar una base de datos con la variable dependiente (ventas semanales) y las variables independientes
Mercadotecnia elaborará un plan desplegar modelo predictivo en fases:
Paso 5. Plan de ejecución.
Mercadotecnia elaborará un plan para desplegar modelo predictivo en fases:
Fase 1. Piloto (San Antonio, TX).
Fase 2. Texas
Fase 3. EUA
Sistemas asegurará la captura del Markdown en las bases de datos.

Creación de la base de datos consolidada

Asegurar “STORES” a “TRAIN” 

bd <- merge(train, stores, by= "Store")

Agregar “FEATURES a”BD” 

bd1 <- bd 
bd1 <- merge(bd1, features)

Eliminar columnas 

bd2 <- bd1
str(bd2)
## 'data.frame':    421570 obs. of  16 variables:
##  $ Store       : int  1 1 1 1 1 1 1 1 1 1 ...
##  $ Date        : chr  "01/04/2011" "01/04/2011" "01/04/2011" "01/04/2011" ...
##  $ IsHoliday   : logi  FALSE FALSE FALSE FALSE FALSE FALSE ...
##  $ Dept        : int  49 26 81 34 59 30 7 85 8 28 ...
##  $ Weekly_Sales: num  13168 5947 28545 9950 317 ...
##  $ Type        : chr  "A" "A" "A" "A" ...
##  $ Size        : int  151315 151315 151315 151315 151315 151315 151315 151315 151315 151315 ...
##  $ Temperature : num  59.2 59.2 59.2 59.2 59.2 ...
##  $ Fuel_Price  : num  3.52 3.52 3.52 3.52 3.52 ...
##  $ MarkDown1   : num  NA NA NA NA NA NA NA NA NA NA ...
##  $ MarkDown2   : num  NA NA NA NA NA NA NA NA NA NA ...
##  $ MarkDown3   : num  NA NA NA NA NA NA NA NA NA NA ...
##  $ MarkDown4   : num  NA NA NA NA NA NA NA NA NA NA ...
##  $ MarkDown5   : num  NA NA NA NA NA NA NA NA NA NA ...
##  $ CPI         : num  215 215 215 215 215 ...
##  $ Unemployment: num  7.68 7.68 7.68 7.68 7.68 ...
    bd2 <- subset(bd2, select = -c (MarkDown1,MarkDown2,MarkDown3,MarkDown4, MarkDown5))
summary(bd2)
##      Store          Date           IsHoliday            Dept      
##  Min.   : 1.0   Length:421570      Mode :logical   Min.   : 1.00  
##  1st Qu.:11.0   Class :character   FALSE:391909    1st Qu.:18.00  
##  Median :22.0   Mode  :character   TRUE :29661     Median :37.00  
##  Mean   :22.2                                      Mean   :44.26  
##  3rd Qu.:33.0                                      3rd Qu.:74.00  
##  Max.   :45.0                                      Max.   :99.00  
##   Weekly_Sales        Type                Size         Temperature    
##  Min.   : -4989   Length:421570      Min.   : 34875   Min.   : -2.06  
##  1st Qu.:  2080   Class :character   1st Qu.: 93638   1st Qu.: 46.68  
##  Median :  7612   Mode  :character   Median :140167   Median : 62.09  
##  Mean   : 15981                      Mean   :136728   Mean   : 60.09  
##  3rd Qu.: 20206                      3rd Qu.:202505   3rd Qu.: 74.28  
##  Max.   :693099                      Max.   :219622   Max.   :100.14  
##    Fuel_Price         CPI         Unemployment   
##  Min.   :2.472   Min.   :126.1   Min.   : 3.879  
##  1st Qu.:2.933   1st Qu.:132.0   1st Qu.: 6.891  
##  Median :3.452   Median :182.3   Median : 7.866  
##  Mean   :3.361   Mean   :171.2   Mean   : 7.960  
##  3rd Qu.:3.738   3rd Qu.:212.4   3rd Qu.: 8.572  
##  Max.   :4.468   Max.   :227.2   Max.   :14.313

Cambiar formato de fecha 

bd2$Date <- as.Date(bd2$Date,format= "%d/%m/%Y")
str(bd2)
## 'data.frame':    421570 obs. of  11 variables:
##  $ Store       : int  1 1 1 1 1 1 1 1 1 1 ...
##  $ Date        : Date, format: "2011-04-01" "2011-04-01" ...
##  $ IsHoliday   : logi  FALSE FALSE FALSE FALSE FALSE FALSE ...
##  $ Dept        : int  49 26 81 34 59 30 7 85 8 28 ...
##  $ Weekly_Sales: num  13168 5947 28545 9950 317 ...
##  $ Type        : chr  "A" "A" "A" "A" ...
##  $ Size        : int  151315 151315 151315 151315 151315 151315 151315 151315 151315 151315 ...
##  $ Temperature : num  59.2 59.2 59.2 59.2 59.2 ...
##  $ Fuel_Price  : num  3.52 3.52 3.52 3.52 3.52 ...
##  $ CPI         : num  215 215 215 215 215 ...
##  $ Unemployment: num  7.68 7.68 7.68 7.68 7.68 ...

Contar ceros y negativos 

#install.packages("wordspace")
library(wordspace)
## Loading required package: Matrix
signcount(bd2$Weekly_Sales)
##    pos   zero    neg 
## 420212     73   1285

Eliminar ventas menores a cero 

bd3 <- bd2
bd3 <- bd3[bd3$Weekly_Sales > 0,]
summary (bd3)
##      Store           Date            IsHoliday            Dept      
##  Min.   : 1.0   Min.   :2010-02-05   Mode :logical   Min.   : 1.00  
##  1st Qu.:11.0   1st Qu.:2010-10-08   FALSE:390652    1st Qu.:18.00  
##  Median :22.0   Median :2011-06-17   TRUE :29560     Median :37.00  
##  Mean   :22.2   Mean   :2011-06-18                   Mean   :44.24  
##  3rd Qu.:33.0   3rd Qu.:2012-02-24                   3rd Qu.:74.00  
##  Max.   :45.0   Max.   :2012-10-26                   Max.   :99.00  
##   Weekly_Sales        Type                Size         Temperature    
##  Min.   :     0   Length:420212      Min.   : 34875   Min.   : -2.06  
##  1st Qu.:  2120   Class :character   1st Qu.: 93638   1st Qu.: 46.68  
##  Median :  7662   Mode  :character   Median :140167   Median : 62.09  
##  Mean   : 16033                      Mean   :136750   Mean   : 60.09  
##  3rd Qu.: 20271                      3rd Qu.:202505   3rd Qu.: 74.28  
##  Max.   :693099                      Max.   :219622   Max.   :100.14  
##    Fuel_Price         CPI         Unemployment   
##  Min.   :2.472   Min.   :126.1   Min.   : 3.879  
##  1st Qu.:2.933   1st Qu.:132.0   1st Qu.: 6.891  
##  Median :3.452   Median :182.4   Median : 7.866  
##  Mean   :3.361   Mean   :171.2   Mean   : 7.960  
##  3rd Qu.:3.738   3rd Qu.:212.4   3rd Qu.: 8.567  
##  Max.   :4.468   Max.   :227.2   Max.   :14.313

Agregar número de la semana 

bd4<- bd3
bd4$week_number <- strftime(bd4$Date, format = "%V")
str(bd4)
## 'data.frame':    420212 obs. of  12 variables:
##  $ Store       : int  1 1 1 1 1 1 1 1 1 1 ...
##  $ Date        : Date, format: "2011-04-01" "2011-04-01" ...
##  $ IsHoliday   : logi  FALSE FALSE FALSE FALSE FALSE FALSE ...
##  $ Dept        : int  49 26 81 34 59 30 7 85 8 28 ...
##  $ Weekly_Sales: num  13168 5947 28545 9950 317 ...
##  $ Type        : chr  "A" "A" "A" "A" ...
##  $ Size        : int  151315 151315 151315 151315 151315 151315 151315 151315 151315 151315 ...
##  $ Temperature : num  59.2 59.2 59.2 59.2 59.2 ...
##  $ Fuel_Price  : num  3.52 3.52 3.52 3.52 3.52 ...
##  $ CPI         : num  215 215 215 215 215 ...
##  $ Unemployment: num  7.68 7.68 7.68 7.68 7.68 ...
##  $ week_number : chr  "13" "13" "13" "13" ...
bd4$week_number <- as.integer(bd4$week_number)
str(bd4)
## 'data.frame':    420212 obs. of  12 variables:
##  $ Store       : int  1 1 1 1 1 1 1 1 1 1 ...
##  $ Date        : Date, format: "2011-04-01" "2011-04-01" ...
##  $ IsHoliday   : logi  FALSE FALSE FALSE FALSE FALSE FALSE ...
##  $ Dept        : int  49 26 81 34 59 30 7 85 8 28 ...
##  $ Weekly_Sales: num  13168 5947 28545 9950 317 ...
##  $ Type        : chr  "A" "A" "A" "A" ...
##  $ Size        : int  151315 151315 151315 151315 151315 151315 151315 151315 151315 151315 ...
##  $ Temperature : num  59.2 59.2 59.2 59.2 59.2 ...
##  $ Fuel_Price  : num  3.52 3.52 3.52 3.52 3.52 ...
##  $ CPI         : num  215 215 215 215 215 ...
##  $ Unemployment: num  7.68 7.68 7.68 7.68 7.68 ...
##  $ week_number : int  13 13 13 13 13 13 13 13 13 13 ...
summary(bd4)
##      Store           Date            IsHoliday            Dept      
##  Min.   : 1.0   Min.   :2010-02-05   Mode :logical   Min.   : 1.00  
##  1st Qu.:11.0   1st Qu.:2010-10-08   FALSE:390652    1st Qu.:18.00  
##  Median :22.0   Median :2011-06-17   TRUE :29560     Median :37.00  
##  Mean   :22.2   Mean   :2011-06-18                   Mean   :44.24  
##  3rd Qu.:33.0   3rd Qu.:2012-02-24                   3rd Qu.:74.00  
##  Max.   :45.0   Max.   :2012-10-26                   Max.   :99.00  
##   Weekly_Sales        Type                Size         Temperature    
##  Min.   :     0   Length:420212      Min.   : 34875   Min.   : -2.06  
##  1st Qu.:  2120   Class :character   1st Qu.: 93638   1st Qu.: 46.68  
##  Median :  7662   Mode  :character   Median :140167   Median : 62.09  
##  Mean   : 16033                      Mean   :136750   Mean   : 60.09  
##  3rd Qu.: 20271                      3rd Qu.:202505   3rd Qu.: 74.28  
##  Max.   :693099                      Max.   :219622   Max.   :100.14  
##    Fuel_Price         CPI         Unemployment     week_number   
##  Min.   :2.472   Min.   :126.1   Min.   : 3.879   Min.   : 1.00  
##  1st Qu.:2.933   1st Qu.:132.0   1st Qu.: 6.891   1st Qu.:14.00  
##  Median :3.452   Median :182.4   Median : 7.866   Median :26.00  
##  Mean   :3.361   Mean   :171.2   Mean   : 7.960   Mean   :25.83  
##  3rd Qu.:3.738   3rd Qu.:212.4   3rd Qu.: 8.567   3rd Qu.:38.00  
##  Max.   :4.468   Max.   :227.2   Max.   :14.313   Max.   :52.00

Separar Año, Mes y Día 

bd5 <- bd4
bd5 <- bd5 %>%
  dplyr::mutate(year= lubridate::year(Date), month = lubridate::month(Date), day = lubridate:: day(Date))

Generar regresión lineal 

regresion <- lm(Weekly_Sales ~ Store + Dept + IsHoliday + Type + Size + Temperature + Fuel_Price + CPI + Unemployment + week_number + year + month + day, data=bd5)
summary(regresion)
## 
## Call:
## lm(formula = Weekly_Sales ~ Store + Dept + IsHoliday + Type + 
##     Size + Temperature + Fuel_Price + CPI + Unemployment + week_number + 
##     year + month + day, data = bd5)
## 
## Residuals:
##    Min     1Q Median     3Q    Max 
## -34331 -12895  -5852   5626 671540 
## 
## Coefficients:
##                 Estimate Std. Error t value Pr(>|t|)    
## (Intercept)    1.110e+06  2.999e+05   3.701 0.000214 ***
## Store         -1.426e+02  3.087e+00 -46.198  < 2e-16 ***
## Dept           1.108e+02  1.097e+00 101.013  < 2e-16 ***
## IsHolidayTRUE  8.511e+02  1.391e+02   6.119 9.45e-10 ***
## TypeB         -3.133e+02  1.078e+02  -2.908 0.003642 ** 
## TypeC          5.836e+03  1.840e+02  31.709  < 2e-16 ***
## Size           9.920e-02  9.584e-04 103.511  < 2e-16 ***
## Temperature    3.701e+00  2.133e+00   1.735 0.082688 .  
## Fuel_Price     4.791e+02  1.480e+02   3.237 0.001207 ** 
## CPI           -2.340e+01  9.996e-01 -23.409  < 2e-16 ***
## Unemployment  -2.538e+02  2.062e+01 -12.308  < 2e-16 ***
## week_number    7.678e+02  4.566e+02   1.682 0.092648 .  
## year          -5.485e+02  1.485e+02  -3.695 0.000220 ***
## month         -3.167e+03  1.988e+03  -1.594 0.111036    
## day           -1.281e+02  6.539e+01  -1.959 0.050115 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 21690 on 420197 degrees of freedom
## Multiple R-squared:  0.08982,    Adjusted R-squared:  0.08979 
## F-statistic:  2962 on 14 and 420197 DF,  p-value: < 2.2e-16

Construir un modelo de predicción 

datos_nuevos <- data.frame(Store=1, Dept=1, IsHoliday= TRUE, Type= "A", Size= 151315, Temperature = 59.17, Fuel_Price = 3.524, CPI= 214.8372, Unemployment = 7.682, week_number =1, year =2012, month = 1, day=1)
predict(regresion,datos_nuevos)
##        1 
## 14667.94

Conclusión

A través del modelo predictivo y la regresión lineal se puede analizar las variables que pueden ser más importantes al momento de determinar el número de ventas semanales o mensuales. Gracias a esto, se pueden generar estrategias para poder evitar pérdidas y seguir manteniendo utilidades dentro de la empresa.

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