## ECON 370 LAB 2:  EXPLORATORY DATA ANALYSIS 
## NAME:  
## DATE:  


# preliminaries ----------------------------------------------------------------

## install packages / load libraries as needed and load libraries 
## you'll need at least tidyverse for R, numpy and pandas for Python
## R users will also want to install the package stargazer

# install.packages("tidyverse")
#install.packages("stargazer")

library(tidyverse)
library(stargazer)

# file path --------------------------------------------------------------------

## ONLY KEEP THE ONE OF THESE THAT IS RELEVANT FOR YOU!

## R users:
## specify your file path by defining yourpath as the path to your working directory for this lab
## leave the rest as it is so that 

username <- Sys.getenv("USERNAME")
pjpath <- paste0("C:\\Users\\", username, "\\Dropbox\\ECON-370\\submissions\\labs\\lab2\\")

yourpath <- "[YOUR FILE PATH]"

mypath <- if_else(username == "pj" | username == "pjaki", pjpath, yourpath)

## Python users:
## specify your file path by defining yourpath as the path to your working directory for this lab
## leave the rest as it is so that 

username = os.getenv("USERNAME")
pjpath = f"C:\\Users\\{username}\\Dropbox\\ECON-370\\submissions\\labs\\lab2\\"
yourpath = f"[YOUR FILE PATH]"

if username == "pj" or username == "pjaki":
  mypath = pjpath
else:
  mypath = yourpath

print(mypath)

# load data --------------------------------------------------------------------

## load the data set ECON370-provisions-transactions.csv
## it contains data on all regular sales transactions at provisions between 
##     2023-01-01 and 2024-06-30


# data cleaning ----------------------------------------------------------------

## based on your table and your exploration of the data, what cleaning would you do?
## should any observations be dropped?
## do any variables need to be modified?


# summary statistics -----------------------------------------------------------

## make a summary statistics table reporting the means, SDs, minima, maxima, and Ns 
##    for all of the numeric variables in the data set.

## Suggestion:  for this lab only, converting the variable names in your summary stats table 
##    to read-able English words only counts for 1 point (out of 100)
##    DON'T WASTE TIME DOING THAT UNTIL YOU'VE FINISHED EVERYTHING ELSE IN THE LAB!

## R hint:  use stargazer! (make sure to convert your tibble to a data.frame first)
## save the table as an 
## set header and float to FALSE if you want to use my standalone latex table template
## Python hint: describe the dataframe and transpose the results

## submit your table as a latex/pdf or html using the templates provided on the web


# visualizing transactions and sales -------------------------------------------

## basic ggplot code to make a histogram + kernel density of transaction amounts
ggplot(clean.prov, aes(amt.total)) + 
  geom_histogram(aes(y = after_stat(density))) +
  geom_density()

## modify the code above to make your graph as close as possible to the one on the course web site
## don't worry to much about matching the colors, just pick something vaguely similar


## save your finished product

ggsave(paste0(mypath, "ECON370-lab2-transactions-hist-YOURNAME.pdf"))

## now make an additional histogram and kernel density where you aggregate transactions 
##    into longer periods (e.g. days, weeks, or months) to show another aspect of the data
## R hint:  look at the R4DS chapter on dates to get quick help on how to aggregate by date
## Is it better to present this information on a log scale or a linear scale?  Why?



# visualizing two variables ----------------------------------------------------

## basic ggplot code to make a scatter plot of weekly gross sales totals and weekly total amounts received (revenues)
##    along with a linear fit (method = "lm") and a non-parametric kernel regression fit

ggplot(weekly.data, aes(x = weekly.gross, y = weekly.total)) + 
  geom_point() +
  geom_smooth() +
  geom_smooth(method = "lm") +
  scale_x_continuous(trans = "log", breaks = c(3500, 7000, 14000, 28000)) +
  scale_y_continuous(trans = "log", breaks = c(3500, 7000, 14000, 28000)) 

## adapt this code to make one of the following more interesting scatter plots:
##    - total sales/revenues within a day/week/month and the # of transactions
##    - comparing sales on two days within the same week (e.g. Friday vs. Saturday)
##    - Comparing sales on Saturday vs. Sunday within the same weekend 
##        (remembering that Sunday is day 1 of the next week)

## make your graph look nice =)