Data visualization
Arnas Matusevičius
2022-02-14
Homework assignment No.1
- Choose a data set (the number of data attributes should be more than 5), explain why it is important or interesting for you.
- Formulate research questions (for which you expect to find the answers)
- Make some visualizations for the formulated questions.
- Prepare a presentation (where you explain the data, questions, problems, results) and upload it.
1. Chosing the dataset
This dataset contains historical information about cargo that has been assigned to various trucks. The dataset contains 66 objects with 16 attributes. An example of the first few objects is presented below.
## ID Value SenderCity ReceiverCity DateCreated
## 1 165212 288 Oberpframmern Vilnius 2021-10-29 15:15:19.560
## 2 165754 1400 Ochtendung Radziunu k. 2021-11-09 12:53:23.840
## 3 165761 590 Lage Moscow 2021-11-09 14:10:58.930
## 4 166085 1060 Bremen Vilnius 2021-11-11 15:54:19.480
## 5 166125 2450 Leinfelden-Echterdingen Ivanovo 2021-11-12 09:50:32.430
## 6 166255 60 Weinstadt Vilnius 2021-11-15 13:44:16.883
## SenderX SenderY ReceiverX ReceiverY LDM Weight Volume UnitTypeID
## 1 48.02556 11.801236 54.63984 25.27138 1.00 456.0 1.800 50
## 2 50.35402 7.423818 54.36220 23.98614 8.50 500.0 42.000 50
## 3 51.98553 8.848820 55.89406 37.44395 0.45 382.0 1.620 50
## 4 53.12085 8.734309 54.63984 25.27138 7.00 9000.0 24.480 33
## 5 48.70013 9.146070 56.99623 41.04270 3.60 4264.0 9.072 36
## 6 48.81519 9.375821 54.63984 25.27138 0.10 9.2 0.097 44
## FirstDimension SecondDimension UnitTypeName
## 1 200 100 PL
## 2 850 200 PL
## 3 118 88 PL
## 4 125 85 EP(120x80x220)
## 5 120 120 PL(120x120x220)
## 6 59 40 BOXI chose this dataset because it is from one of the projects I’m currently working on. This data so far has been used only to depict truck routes, hence I hope to obtain a more in-depth look into what causes various decisions to be made during route planning.
2. Research questions
- Do the dimensional characteristics, namely LDM (loading meter), weight and volume, influence the value of cargo?
- What is the distribution of value among different cargo types?
- Do certain regions have better paying cargo than others?
3. Visualizations
corrplot(cor(data.frame(Value = data[,2]),data[,c(10:12)]), type = "upper", diag = T, method = "pie", title = "The correlation between value and other attributes")- The calculated correlation coefficients show that the Value depends most on the LDM attribute of cargo. In general, higher-dimensional attributes bring more value to cargo.
p <- ggplot(data, aes(x = UnitTypeName, y = Value, fill = UnitTypeName)) +
ggtitle("Values for different types of cargo") +
geom_boxplot(outlier.colour="black",outlier.shape=16,outlier.size=3, notch=F) +
labs(x = "", fill = "Cargo type")- The highest valued cargo is observed to be of PL(120x120x220) type. Since it is only a single occurrence, the PL(120x100x220) type is identified as the highest valued cargo type with the median value of around 750.
The rest of the cargo is distributed significantly lower around the 250 mark with a few outliers.
g = list(
scope = "europe",
domain = list(
x = c(0, 1),
y = c(0, 1)
),
center = list("lat" = 50.826941, "lon" = 10.4261393),
lataxis = list(range = c(48.3, 55.5)),
lonaxis = list(range = c(5.9, 14.9)),
showland = TRUE,
landcolor = "rgb(229,236,246)",
showframe = TRUE,
projection = list(type = "Mercator"),
resolution = 50,
countrycolor = "rgb(102,51,153)",
coastlinecolor = "rgb(102,51,153)",
showcoastlines = TRUE
)
fig <- plot_geo(data, lat = ~SenderX, lon = ~SenderY, color = ~Value, size = ~Value)
fig <- fig %>% add_markers(
text = ~paste(data$Value, "eu"), hoverinfo = "text", marker = list(sizeref=0.09, sizemode="area")
)
fig <- fig %>% layout(title = 'Distribution of cargo by value', geo = g)- The cargo seems to be clustered to the west side of Germany. However, clear clustering of high valued cargo is not observed, as such cargo appears to be evenly spread out amongst all cargo.