USA Investment Analysis with R

It’s been a long time since I’ve participated in #TidyTuesday: a bit due to lack of time, a bit because the datasets they published didn’t motivate me, and a bit because I was embarrassed to always upload the same thing. But today, after having dedicated three hours of my Sunday to Kierisi’s livestream on Machine Learning, I wanted to continue programming.

(For those who don’t know: #TidyTuesday is a collaborative project that consists of working every week with a set of real data to practice visualization and other skills in the world of data science. You can find more information in its GitHub repository and under the hashtag on twitter.)

Let’s start with the general organization:

library(tidytuesdayR)
library(tidyverse)
library(hrbrthemes)
library(streamgraph)
library(htmlwidgets)
library(webshot)

tuesdata <- tidytuesdayR::tt_load('2021-08-10')

## 
## 	Downloading file 1 of 3: `ipd.csv`
## 	Downloading file 2 of 3: `chain_investment.csv`
## 	Downloading file 3 of 3: `investment.csv`

investment <- tuesdata$investment
chain_investment <- tuesdata$chain_investment
ipd <- tuesdata$ipd


# Exploratory Analysis ----------------------------------------------------

glimpse(investment)

## Rows: 6,106
## Columns: 5
## $ category  <chr> "Total basic infrastructure", "Total basic infrastructure", …
## $ meta_cat  <chr> "Total infrastructure", "Total infrastructure", "Total infra…
## $ group_num <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, …
## $ year      <dbl> 1947, 1948, 1949, 1950, 1951, 1952, 1953, 1954, 1955, 1956, …
## $ gross_inv <dbl> 4974.662, 6486.770, 7376.143, 7943.959, 8961.233, 9707.590, …

glimpse(chain_investment) # how much did it change with time

## Rows: 6,035
## Columns: 5
## $ category        <chr> "Total basic infrastructure", "Total basic infrastruct…
## $ meta_cat        <chr> "Total infrastructure", "Total infrastructure", "Total…
## $ group_num       <dbl> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, …
## $ year            <dbl> 1947, 1948, 1949, 1950, 1951, 1952, 1953, 1954, 1955, …
## $ gross_inv_chain <dbl> 73278.62, 83218.25, 95760.00, 103642.40, 102264.39, 10…

glimpse(ipd)

## Rows: 6,106
## Columns: 5
## $ category      <chr> "GDP", "GDP", "GDP", "GDP", "GDP", "GDP", "GDP", "GDP", …
## $ meta_cat      <chr> "GDP", "GDP", "GDP", "GDP", "GDP", "GDP", "GDP", "GDP", …
## $ group_num     <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ year          <dbl> 1947, 1948, 1949, 1950, 1951, 1952, 1953, 1954, 1955, 19…
## $ gross_inv_ipd <dbl> 12.250, 12.946, 12.942, 13.064, 13.950, 14.254, 14.436, …

investment %>% group_by(category) %>% count()

## # A tibble: 60 × 2
## # Groups:   category [60]
##    category                                n
##    <chr>                               <int>
##  1 Air transportation                     71
##  2 All federal                            71
##  3 Conservation and development           71
##  4 Education                              71
##  5 Federal                               639
##  6 Federal electric power structures      71
##  7 Health                                 71
##  8 Highways and streets                   71
##  9 Office buildings, NAICS 518 and 519    71
## 10 Other federal                          71
## # ℹ 50 more rows

investment %>% group_by(meta_cat) %>% count()

## # A tibble: 16 × 2
## # Groups:   meta_cat [16]
##    meta_cat                             n
##    <chr>                            <int>
##  1 Air /water /other transportation   852
##  2 Conservation and development       213
##  3 Digital                            355
##  4 Education                          426
##  5 Electric power                     426
##  6 Health                             497
##  7 Highways and streets               142
##  8 Natural gas /petroleum power       213
##  9 Power                              213
## 10 Public safety                      355
## 11 Sewer and waste                    142
## 12 Social                             426
## 13 Total basic infrastructure         568
## 14 Total infrastructure               426
## 15 Transportation                     710
## 16 Water supply                       142

This week’s dataset was about investments in the US during the period of time that goes from 1947 to 2017. I have to confess that I wasn’t very motivated by the topic either; in addition, the three dataframes that they brought were very similar and with some terminology about economics that I did not fully understand. But it did feel like doing something, and that was enough.

I decided not to procrastinate too much and make at least two charts on some familiar field, so I chose education.

investment %>% 
  filter(category=="Federal") %>% 
  mutate(highligth = ifelse(meta_cat=="Education", "Education", "Other")) %>% 
  group_by(meta_cat, year) %>% 
  ggplot(aes(x = year, y = gross_inv, group = meta_cat, size = highligth))+
  geom_line()+
  scale_color_manual(values = c("#69b3a2", "lightgrey"))+
  scale_size_manual(values=c(1.5,0.2))+
  theme_ipsum()+
  theme(legend.position = "none")+
  labs(x="", y="",
       title="Evolution of Federal Investment in Education",
       caption = "Data: Bureau of Economic Analysis | Plot: @_msquiroga | #TidyTuesday")

## Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
## ℹ Please use `linewidth` instead.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.

graph <- investment %>% 
  filter(meta_cat == "Education") %>% 
  streamgraph(., key = "category", value = "gross_inv", date="year") %>% 
  sg_legend(TRUE, "Fuente de inversión")
graph