Introduction

Our goal today is to educate you on the breakdown of spending within all age groups at Regork and draw insights on how to market and draw in these customers to continue shopping at our stores.

We are looking to analyze the monthly amount sold for the top 3 categories for each age group identified by Regork. Once we have this data we will be able to pull out the down months and counter those with sales or coupons to drive sales throughout the year. You will be seeing tables and graphs to help draw attention to areas of extreme concern.

Overall, we know that it is important to bring customers into the stores each week to have consistent sales throughout the year. This is why we have spent time bringing you digestible insights on how to increase our sales and customer volume consistently for the sustainable future.

Our methodology was to combine the transactions, products, and demographics data together and sort that data into groups based on age groups. Once we had them grouped, we were able to sort each via their highest bought product categories. We then broke that data down by months to get the yearly view on the data. This allowed us to see months that the product was selling great and months where the product needed some support. From there, we pulled our final insights and will present those to you at the end of this report.

Data Used

During our analysis, we used the completejourney data set. This data set is comprised of household level transactions over one year time span from a group of approximately 2,500 households who are frequent shoppers at a grocery store. We used a combination transactions, demographics, and products data to come to our final conclusions and make sure our work was exhaustive and complete.

Here are the individual packages we used and their essential functions:

dplyr - Primarily a set of functions designed to enable dataframe manipulation

tidyverse - A collection of open source packages for the R programming language that share an underlying design philosophy, grammar, and data structures” of tidy data

lubridate - Functions to work with date-times and time-spans

ggthemes - Provides extra themes, geoms, and scales for the ggplot2 package

ggplot2 - A system for declarative creating graphics, based on The Grammar of Graphics

paged/rmarkdown - For building table styles

library(completejourney)
library(dplyr)
library(tidyverse)
library(lubridate)
library(ggthemes)
library(ggplot2)
library(rmarkdown)
transactions <- get_transactions()

Data Preparation & Tidy Data

As mentioned previously, we used the transactions, products, and demographics data. Within that we used the columns product_category(to group by top bought categories), age, transaction_timestamp(to filter by month), and quantity(this played into the product_category data). We loaded this in and the packages mentioned in the last section via the library() function before proceeding to our analysis.

This is how we joined our data

transactions %>%
 left_join(products) %>%
  left_join(demographics) -> df_main

We started by filtering per month for later analysis

transactions %>%
  mutate(month = month(transaction_timestamp)) -> monthly_data

This is an example of how we broke the overall data into each age group. We repeated this as needed.

df_main %>%
  filter(str_detect(age, "19-24") ) -> df_19to24

This is an example of how we broke down from filtering, slicing, and eventually graphing an output. We did this to find the top 3 of each age group and graphed each category individually. This resulted in 18 graphs which you will see in the next section.

df_19to24 %>%
  group_by(product_category) %>% # grouping by category  
  
  count() %>%                    # finding top bought categories  
  
  arrange(desc(n)) %>%
  filter(n>1440) %>%             # filtering for only the top 3 results  
  
 left_join(df_19to24)%>%         # rejoining old data to bring back months  
 
  select(product_category, n, transaction_timestamp) %>%
  filter(product_category == "SOFT DRINKS") %>%  # only selecting one of the top 3 at a time  
  
  mutate(only_months = month(transaction_timestamp)) %>%
  group_by(only_months) %>%
  count() %>%                    # breaking it down by each month to then graph  
  
  ggplot(aes(x=only_months, y = n)) +
  geom_col()                     # baseline graph of a top category by month

For simplicity, we only included one code chunk, but we reworked the above for each age and category until all graphs were finalized. You will see full code for each graph in the Data Analysis section.

Data Analysis

Below you will see tabs for each of the age groups followed by the top three categories of each and the monthly graphical breakdown. Moreover, you can also see a table that lists all age groups and their top three for ease of visualization. All insights will be stated in the summary section.