Exam 1
Emma Fields
2024-10-01
Instructions:
- This exam is open note, though you will want to use a cheat sheet to answer the following questions efficiently.
- The data provided includes the estimated and actual costs of large road construction projects managed by the US Department of Transportation (DoT) in 2013.
- Note, that the numbers provided in this spreadsheet are in thousands (e.g., 100 = $100,000).
Importing and Summarizing Data
- Import the data from “Road construction bids.csv”.
library(readxl)
Road= read.csv("C:\\Users\\etfie\\OneDrive\\Data Visualization\\Road construction bids.csv")- How many bids are in this data set?
There are 235 bids in this data set.
- What are the average estimate and average actual costs of these projects? What can you conclude?
library(dplyr)##
## Attaching package: 'dplyr'## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, unionsummarize(Road, mean_ActualCost = mean(Actual_Cost, na.rm = TRUE),
mean_DoTEstimate = mean(DoT_Estimate, na.rm = TRUE))## mean_ActualCost mean_DoTEstimate
## 1 1268.715 1347.077Manipulating Data
- Add the following variables to your data set:
- difference = the difference between the estimate and the actual cost.
- percent_difference = difference/cost
- budget = projects whose actual cost is more than the estimate are labeled as “Over Budget”, otherwise “Under Budget”.
library(dplyr)
summarize(Road,
mean_ActualCost = mean(Actual_Cost, na.rm = TRUE),
mean_DoTEstimate = mean(DoT_Estimate, na.rm = TRUE))## mean_ActualCost mean_DoTEstimate
## 1 1268.715 1347.077difference
Road$difference <- Road$DoT_Estimate - Road$Actual_Costpercent_difference = difference/cost
Road$percent_difference <- (Road$difference/Road$Actual_Cost)*100budget = projects whose actual cost is more than the estimate are labeled as “Over Budget”, otherwise “Under Budget”.
Road$budget <- ifelse(Road$Actual_Cost > Road$DoT_Estimate, "Over Budget", "Under Budget")large and small projects
Road$Project_Size <- ifelse(Road$Actual_Cost > 1000, "Large", "Small")- What is the mean difference between large and small projects? What does this tell us?
library(dplyr)
LargeProjects <- group_by(Road, Project_Size == "Large")
SmallProjects <- group_by(Road, Project_Size == "Small")
Mean_Large <- summarize(LargeProjects, Mean_Large = mean(difference, na.rm = TRUE))
Mean_Small <- summarize(SmallProjects, Mean_Small = mean(difference, na.rm = TRUE))
Mean_Large## # A tibble: 2 × 2
## `Project_Size == "Large"` Mean_Large
## <lgl> <dbl>
## 1 FALSE 29.8
## 2 TRUE 184.- What is the mean percent difference between large and small projects? What does this tell us?
Percent_Mean_Large <-summarize(LargeProjects, Mean_Large = mean(percent_difference, na.rm = TRUE))
Percent_Mean_Small <- summarize(SmallProjects, Mean_Small = mean(percent_difference, na.rm = TRUE))
PDBLSP = Percent_Mean_Large - Percent_Mean_Small
PDBLSP## Project_Size == "Large" Mean_Large
## 1 0 5.844409
## 2 0 -5.844409Percent_Mean_Large## # A tibble: 2 × 2
## `Project_Size == "Large"` Mean_Large
## <lgl> <dbl>
## 1 FALSE 10.4
## 2 TRUE 4.58Percent_Mean_Small## # A tibble: 2 × 2
## `Project_Size == "Small"` Mean_Small
## <lgl> <dbl>
## 1 FALSE 4.58
## 2 TRUE 10.4Plotting Data
- Plot a graph of the difference (between estimate and actual cost) vs. actual costs. Make your graph as descriptive as possible, adding smoothing, color grouping, scales, and labels to best present the data.
library(ggplot2)
library(dplyr)
ggplot(Road, aes(x = Actual_Cost, y = difference, color = Project_Size)) +
geom_point(alpha = 0.6, size = 3) +
geom_smooth(method = "lm", se = TRUE, linetype = "dashed", color = "black") +
scale_color_manual(values = c("Large" = "orchid", "Small" = "sienna")) +
labs( title = "Difference vs. Actual Cost",
subtitle = "Difference Compared to Actual Cost",
x = "Actual Cost",
y = "Difference (Estimate - Actual Cost)",
color = "Project Size") +
theme_minimal() +
theme(
plot.title = element_text(hjust = 0.5, face = "bold"),
plot.subtitle = element_text(hjust = 0.5)
) +
scale_x_continuous(labels = scales::comma) +
scale_y_continuous(labels = scales::comma) ## `geom_smooth()` using formula = 'y ~ x'
- Plot a graph of the percent difference (between estimate and actual cost) vs. actual costs. Make your graph as descriptive as possible, adding color, grouping, scales, and labels to best present the data.
library(ggplot2)
ggplot(Road, aes(x = percent_difference, fill = Project_Size)) +
geom_histogram(position = "dodge", bins = 30, alpha = 0.7, color = "black") +
scale_fill_manual(values = c("Large" = "violetred", "Small" = "yellowgreen")) +
labs(title = "Percent Difference vs Actual Cost",
subtitle = "Grouped by Project Size",
x = "Percent Difference %",
y = "Frequency",
fill = "Project Size") +
theme_minimal() +
theme(plot.title = element_text(hjust = 0.5, face = "bold"),
plot.subtitle = element_text(hjust = 0.5)) +
scale_x_continuous(labels = scales::percent_format(scale = 1))
- Summarize what these graphs show us.
In the graph on question 7 I noticed that the larger projects costed more that the smaller projects (as expected). I also noticed that the smaller projects seemed to be more frequently in the same part of the graph showing that smaller projects are more consistent between their estimated cost-actual cost and their actual cost, whereas the larger projects were not as consistent with their difference and actual cost.
In the graph on question 8 I noticed that the larger projects had more of a percent difference between their estimated cost and their actual cost than the smaller projects did. The larger projects have a larger frequency, however, it seems that the smaller projects don’t have as big of a percent difference as the larger projects do.