retailDataAnalysis.Rmd

---
title: "Retail Data Analysis"
author: "Deo Ivan Mareza"
date: "1/16/2020"
output: 
  html_document:
    toc: true
    toc_depth: 4
    number_section: true
    toc_float:
      collapsed: true
    highlight: breezedark
    theme: cerulean
    
---

```{r setup, include=FALSE}
library(GGally)
library(dplyr)


knitr::opts_chunk$set(echo = TRUE)
```

# Brief Summary

Hello and welcome to my training R document. In this document, I will try to analyze a retail data of a company, and see what kind of conclusion and business insight we might have.

The main objective of this document is to practise and sharpen my R command, as well compare data and get some insights on the given data.

# The Data

```{r}
# data set

retail <- read.csv(file="data_input/retail.csv")


```


This is the first look of the data that we received.

The only cleanup we need is to change some factors to characters, and removing some columns that we don't need.



```{r echo=TRUE}

head(retail)

```

# Data Cleanup

At this stage, there's a few things that I'd like to do, which is : 
*1. Remove Row.ID, Order.ID, and Customer.ID*
*2. Fix data type*

```{r echo=TRUE}

# retail <- read.csv(file="data_input/retail.csv")


colnames(retail)

retail <- retail[,-c(1,2,6)]
# retail <- retail[,-"Customer.ID"]

```

```{r echo=TRUE}

retail$Order.Date <- as.Date(retail$Order.Date, "%m/%d/%y")
retail$Ship.Date <- as.Date(retail$Ship.Date, "%m/%d/%y")


retail$Product.ID <- as.character(retail$Product.ID)
retail$Product.Name <- as.factor(retail$Product.Name)


str(retail)

```

Now that we have removed the columns we do not need as well as making sure all the data types are correct, we're ready to begin the analysis. 



# Data Analysis

The data provided is a sales data within the period of 3rd January 2014 to 30th December 2017.

```{r echo = TRUE}

min(retail$Order.Date)
max(retail$Order.Date)


```

## TOP 10



**Top 10 Products based on Quantity :**


```{r echo = TRUE}

productSoldQty <- aggregate(formula = Quantity ~ Product.Name, data = retail, FUN = max)

productSoldQty[order(productSoldQty$Quantity, decreasing = T),][1:10,]

```


**Top 10 Most Profitable Product :**



Using aggregates to count the max profit


```{r}

agg1 <- aggregate(formula = Profit ~ Product.Name, data = retail, FUN = max )

agg1[order(agg1$Profit, decreasing=TRUE),][1:10,]


```


In addition, Canon imageClass 2200 is sold for 20 units.

```{r echo=TRUE}

canonImageClass <- retail[grep("Canon imageCLASS 2200", retail$Product.Name),]

sum(canonImageClass$Quantity)


```




**Top 10 Most UNProfitable Product :**


```{r echo=TRUE}

agg2 <- aggregate(formula = Profit ~ Product.Name, data = retail, FUN = min )

agg2[order(agg2$Profit),][1:10,]


```


## Most Profitable Segment


**A comparison on Profit and Quantity Sold based on consumer segments** 

```{r}

agg3 <-  aggregate(formula = cbind(Profit, Quantity) ~ Segment, data = retail, FUN = sum)

agg3[order(agg3$Profit, decreasing = T),]


```





## Preferrable Shipping Mode based on Segment


Total of Shipping Mode chosen : 

```{r echo=T}

# table(retail$Ship.Mode, retail$Segment)

agg4 <- aggregate (formula = Quantity ~ Ship.Mode, data = retail, FUN = sum)

agg4[order(agg4$Quantity, decreasing = T),]

```

Shipping Mode based on Segment : 

```{r}

table(retail$Ship.Mode, retail$Segment)

```



## Shipping Profitability


**Analyzing different shipping options profitability**

```{r echo=T}

correlation2 <- cbind.data.frame(retail$Ship.Mode, retail$Profit)

# ggcorr(data = correlation2, label = 1)
aggregate(formula = Quantity ~ Ship.Mode, data = retail, FUN = sum)
aggregate(formula = Profit ~ Ship.Mode, data = retail, FUN = sum)

ggpairs(retail, columns = c("Ship.Mode", "Profit"),ggplot2::aes(colour="species"))

```



## Discount Effectiveness


**Just how effective is the discount promotion?**

```{r echo = TRUE}

cor(retail$Discount, retail$Sales)
cor(retail$Discount,retail$Profit)

```


**Correlation on Discount, Sales, Quantity, and Profits**

```{r echo=TRUE}



correlation1 <- cbind.data.frame(retail$Discount, retail$Sales, retail$Quantity, retail$Profit)

ggcorr(data = correlation1, label = 1)

```

```{r echo=T}


ggpairs(retail, columns = c("Discount", "Profit"))

```

```{r}



```





## Sales and Profit Scatterplot


**Correlation on Sales and Profit**

```{r echo=T}


ggpairs(retail, columns = c("Sales", "Profit"))


```



# Data Conclusion

There's a few things that we can conclude from the short analysis above :
  *The business should consider to stop selling Cubify CubeX 3D Printer Double Head Print
  *Based on correlation analysis between discount to sales as well as discount to profits, it doesn't seem that the discount is effective in increasing sales.
  *After analyzing Sales and Profit scatter plot, it seemed that there's a cluster in a low performing number, where sales and profit are close to 0. The company should revise their pricing strategy.
  *It seemed that the standard class shipping mode has the biggest profit, as well as the preferred shipping mode for a lot of the customers.
  *Consumer segment seemd to yield most profit and the most sale. This may be a good insight for the marketing or sales team.