Data Manipulation and Visualization Techniques in Julia – Part I

In this article, we’ll look at data manipulation and visualization techniques in Julia. However, I’ll not get into the details of each parameter of every function, as the objective of this series is to use Julia as a tool to achieve our goal, i.e. building and backtesting trading strategies. So, we’ll stay focused on that.

You can refer to the detailed documentation of a function if you need it to solve any particular challenge you face while programming.

This article is divided into the following sections:

• Data manipulation
  • Creating new dataframes
  • Indexing and summarising data
  • Basic mathematical operations
  • General operations
  • Grouping data
  • Dealing with missing data
• Importing and exporting data as CSV and excel files
• Data visualization
  • Line plot
  • Attributes of a plot
  • Scatter plot
  • Heatmap
  • Histogram
  • Pie chart
  • Plotting mathematical functions
  • Saving plots
  • Animated plots
  • Various packages for plotting in Julia

In my previous posts in this Julia programming series, I introduced the language and started with the basic syntax of Julia programming. You can check that out as well.

---

Data manipulation

You need to understand the data structures dealing with large heterogeneous data sets whenever you work with any programming language. In the Julia world, they are called dataframes.

Julia’s DataFrames.jl package provides a way to structure and manipulate data.

It can be installed using the “Pkg” module.

Creating new dataframes

Here’s an example of creating a new dataframe.

## Creating a new dataframe
Using Pkg
Pkg.add(“DataFrames”)
Using DataFrames
df_1 = DataFrame(Name = ["Vivek","Viraj","Rohan","Ishan"], 
               Team = ["EPAT", "Marketing", "Sales", "Quantra"],
               Work_experience = [15, 8, 7, 10]
               )

## Creating another dataframe

df_2 = DataFrame(a=rand(10), b=rand(10))

Creating a new dataframe.jl hosted with ❤ by GitHub

Output:

Name	Team	Work_experience
String	String	Int64
Vivek	EPAT	15
Viraj	Marketing	8
Rohan	Sales	7
Ishan	Quantra	10

a	b
Float64	Float64
0.845011	0.720306
0.647665	0.0409036
0.427267	0.221369
0.413642	0.374832
0.477994	0.118461
0.0849006	0.157679
0.0477405	0.845332
0.518909	0.159305
0.93499	0.259579
0.60034	0.115911

Column names can be accessed using the names() function.

# Accessing column 

namesnames(df_1) #As symbol

propertynames(df_1) # As names

accessing columns.jl hosted with ❤ by GitHub

Output:

3-element Vector{String}:
"Name"
"Team"
"Work_experience"

3-element Vector{Symbol}:
:Name
:Team
:Work_experience

Renaming columns can be done using the rename() function.

rename!(df_1, ["name", "team", "work experience"])

rename columns.jl hosted with ❤ by GitHub

name	team	work experience
String	String	Int64
Vivek	EPAT	15
Viraj	Marketing	8
Rohan	Sales	7
Ishan	Quantra	10

Indexing and summarising data

Indexing dataframes to use particular rows or columns for manipulation is a fundamental operation, and summarising data helps us understand it better. In Julia, summary stats of any dataset can be printed using the describe() function.

describe(df_2)

describe data.jl hosted with ❤ by GitHub

variable	mean	min	median	max	nmissing	eltype
Symbol	Float64	Float64	Float64	Float64	Int64	DataType
a	0.499846	0.0477405	0.498452	0.93499	0	Float64
b	0.301368	0.0409036	0.190337	0.845332	0	Float64

Another way to find the number of rows and columns in a dataframe is using ncol() and nrow() functions.

## No. of columns and rows in a dataframe
ncol(df)
nrow(df)

row and columns.jl hosted with ❤ by GitHub

Output:
2
10

Let’s look at multiple methods of accessing rows and columns of a dataframe.

# Methods to access columns and rows in a dataframe

df_1.name

df_1."team"

df_1[1:3, "team"]

df_1[1, :]

df_1[!, 1:2]

methods of accessing columns and rows.jl hosted with ❤ by GitHub

Output:
4-element Vector{String}:
"Vivek"
"Viraj"
"Rohan"
"Ishan"

4-element Vector{String}:
"EPAT"
"Marketing"
"Sales"
"Quantra"

3-element Vector{String}:
"EPAT"
"Marketing"
"Sales"

name	team	work experience
String	String	Int64
Vivek	EPAT	15

name	team
String	String
Vivek	EPAT
Viraj	Marketing
Rohan	Sales
Ishan	Quantra

Stay tuned for the next installment in which Anshul Tayal will discuss basic mathematical operations.

Visit QuantInsti to read the full article: https://blog.quantinsti.com/data-manipulation-visualization-using-julia/.