Manipulaton of Time Series Data

In R, there are special time series objects – called as Date-time objects which are designed to work best with many time series statistical models and visualisation libraries such as ggplot2.

These date time objects ensure that each observations time and value are "tied toether" so that they are not easily separated during analysis.

It is very common to have the time component (date,time etc.) of the time series data as Character object and not as date-time object. R has special classes called as Date and POSIXct class, that are designed to handle time series data. These classes have special methods that work very efficiently with time series data. R will only recognize the objects belonging to these clasees as date-time objects.

In this video, we will learn how to convert the character class objects into Date/POSIXct class objects using various parsing functions from lubridate package. We will use stock pricing data and IoT sensor data. These data-sets have time component (date variable) class as character class. We will convert this class into either Date or POSIXct class using parsing functions.

Datasets and Rmarkdown file can be found at Git repo:
https://github.com/siddharth-sahasrabudhe/time-series-data

1 0

We have a time series data, in this video we will extend our discussion further and look at various functions from lubridate package that are used to extract the information from date-time objects. For e.g. if I want to extract only year from date object, or only month from date object..I can able to do this using various extract functions.

In this video, we will carry out exploratory data analysis of flights data from nycflights13 package. We will filter out Delta Airways and try to answer below questions:
On average, Which month has the largest arrival delay?
On average, Which day of a week has the largest arrival delay?
On average, Which hour of a day has the largest arrival delay?

We will start with month and drill down to day and hour. We will use various functions from lubridate package to extract out the month, day and hour information from time-series data.

The Rmarkdown file can be found at Git Repo:
https://github.com/siddharth-sahasrabudhe/time-series-data/blob/main/extract_information.Rmd

The dplyr package video series can be found at:
https://youtu.be/Kxt5H0LM-1A – How to use mutate function
https://youtu.be/JMHJReB00ZA – Generate data summary using group_by ( ) and summarise ( ) functions

ggplot2 package video series can be found at:
https://youtu.be/wUWwL5ksIXM – How to create a bar chart

Link to previous video on lubridate package is:
https://youtu.be/LBh1QS-4Og0- Parsig time-series data.

1 0

In this video we will learn the basic arithmetic operations that can be performed on date time objects. We will also look at the time span concept – lubridate package has two time span classes :

– Duration class
– Period class

Duration – The Duration class will always supply mathematically precise results. A duration year will always equal 365 days. Duration class follows stop-watch concept of time-span. Thus, duration class won't take into account leap years or any other time changes.

Period – Periods, on the other hand, fluctuate the same way the timeline does to give intuitive results.
Period class follows human concept of time span. Periods will take care of leap year, day-light saving calculations and generate the result.

We will work on share_price and transactions data set and try to use arithmetic operators on its date-time variables.

Video link on Parsing Date time objects:
https://youtu.be/LBh1QS-4Og0

Video link on filter function from dplyr package
https://youtu.be/wAQGj5w4xy4

Video link on mutate function from dplyr package
https://youtu.be/Kxt5H0LM-1A

Video link on summarise function from dplyr package:
https://youtu.be/JMHJReB00ZA

The Rmarkdown file with data-sets is available on Github repo:
https://github.com/siddharth-sahasrabudhe/time-series-data

0 0

Interval is a duration with specific start and end date. lubridate package comes with helper functions that work with interval.

– You can retrieve the amount of time interval covers, either as duration or period.

– You can test whether another date time is in the interval

– You can test whether two intervals overlap.
In this video, we will work with various helper functions that make working with time interval easy. We will use Indian Prime Minister's data set and use various time interval helper functions on it. We will try to answer some interesting questions such as, Who was the Prime Minister at office when some of the major events have unfolded in India's history post independence era.

The video link for parsing functions in lubridate package is:
https://youtu.be/LBh1QS-4Og0

The video link for mutate function in dply package is:
https://youtu.be/Kxt5H0LM-1A

Rmarkdown file and data set can be found at Git repo:
https://github.com/siddharth-sahasrabudhe/time-series-data

0 0