2025-11-07
Statistics have numerous tools to work with data. R has access to specific packages that assist in data analysis, in the case of this presentation, the dplyr package and its associated functions and its integration with dataframes.
dplyr’s packages can alter dataframes in many different ways, allowing for refined and exact statistical analysis to be performed when all are used in conjunction with each other.
## Length Width Height ## 1 8 3 20 ## 2 23 6 25 ## 3 3 10 23 ## 4 15 6 19 ## 5 11 9 23 ## 6 11 7 26 ## 7 19 6 25 ## 8 23 6 19 ## 9 10 4 20 ## 10 20 5 22
Boxes is a 3 column, 10 row dataframe that describes the dimensions of fictional 10 boxes, which will prove useful for displaying dplyr functions, what they do, and how they can prove beneficial in statistical analysis.
This 3d plot showcases the unaltered version of the Boxes dataframe that will be used in this presentation. This dataframe will be altered in different ways with each concept.
Dplyr includes a number of functions in its library, each beneficial in different ways for dataframe analysis and manipulation. Notable dplyr functions are listed below:
filter() - Extraction of specific rowsarrange() - Ordering and sorting rowsselect() - Extraction of specific columnsmutate() - Create new columns, using existing or new datasummarize() - Summarize many values into one locationgroup_by() - Change scope of other functionsDplyr functions often often tibbles, a modernized version of dataframes. This presentation will not describe the differences between them and will refer to both tibbles and dataframes as dataframes
The filter() function can filter a dataframe to only include rows that are within the specified filter. The example dataframe, Boxes, has 3 columns and 10 rows, Length, Width, and Height, each with 10 values.
filter() is told what dataframe to use, and then given criteria to search for only rows with Height \(\le 23\) and Width \(> 6\).
filter(Boxes, Height <= 23, Width > 6)
## Length Width Height ## 1 3 10 23 ## 2 11 9 23
arrange() can sort columns in ascending or descending order, allowing for the largest and smallest values of a column to be quickly found for further analysis.
arrange(Boxes, Width)
## Length Width Height ## 1 8 3 20 ## 2 10 4 20 ## 3 20 5 22 ## 4 23 6 25 ## 5 15 6 19 ## 6 19 6 25 ## 7 23 6 19 ## 8 11 7 26 ## 9 11 9 23 ## 10 3 10 23
select() allows for dataframe to be trimmed to only include specific columns or specific ranges of columns, allowing for quick separation of data for use elsewhere in different contexts. Additionally, the trimmed columns can be defined as their own dataframes.
select(Boxes, Length, Width)
## Length Width ## 1 8 3 ## 2 23 6 ## 3 3 10 ## 4 15 6 ## 5 11 9 ## 6 11 7 ## 7 19 6 ## 8 23 6 ## 9 10 4 ## 10 20 5
This graph plots the data from a newly created dataframe using the parameters from the previous slide, trimmed with select().
mutate() allows for the creation of new columns on a dataframe using existing columns as data for operations, or for the implementation of entirely original columns.
mutate(Boxes, Volume = Length*Width*Height, Color="Purple")
## Length Width Height Volume Color ## 1 8 3 20 480 Purple ## 2 23 6 25 3450 Purple ## 3 3 10 23 690 Purple ## 4 15 6 19 1710 Purple ## 5 11 9 23 2277 Purple ## 6 11 7 26 2002 Purple ## 7 19 6 25 2850 Purple ## 8 23 6 19 2622 Purple ## 9 10 4 20 800 Purple ## 10 20 5 22 2200 Purple
Complex operations can be performed to create new columns using mutate().
Ex: Creation of a dataframe with values defined by \(\displaystyle{5\pi \over 23} \times Height = NewColumn\).
## Length Width Height NewColumn ## 1 8 3 20 13.65910 ## 2 23 6 25 17.07387 ## 3 3 10 23 15.70796 ## 4 15 6 19 12.97614 ## 5 11 9 23 15.70796 ## 6 11 7 26 17.75683 ## 7 19 6 25 17.07387 ## 8 23 6 19 12.97614 ## 9 10 4 20 13.65910 ## 10 20 5 22 15.02501
Using the Boxes dataframe assume that the units are in meters. A conversion of each column into feet could be done in a manner such as
\(Column(m) \times 3.281 = NewColumn(ft)\).
Doing so for each relevant column results in the following dataframe:
## Length Width Height meterLength meterWidth meterHeight ## 1 8 3 20 26.248 9.843 65.620 ## 2 23 6 25 75.463 19.686 82.025 ## 3 3 10 23 9.843 32.810 75.463 ## 4 15 6 19 49.215 19.686 62.339 ## 5 11 9 23 36.091 29.529 75.463 ## 6 11 7 26 36.091 22.967 85.306 ## 7 19 6 25 62.339 19.686 82.025 ## 8 23 6 19 75.463 19.686 62.339 ## 9 10 4 20 32.810 13.124 65.620 ## 10 20 5 22 65.620 16.405 72.182
summarize() collapses a dataframe into one row, allowing review and analysis of the overall data following the use of a function over an entire column or groups in a column, collapsed into 1 result or group of results. This example takes the mean of all values across the entire dataframe for Length and yields a 1 row dataframe, as mean only results in 1 answer in this case when it is the mean for the whole dataframe.
summarize(Boxes, AvgLength = mean(Length, na.rm = TRUE))
## AvgLength ## 1 14.3
group_by() redefines the scope of other functions. In this example, summarize() is used on a new scope of data defined by group_by() to obtain the mean Length for each group of Width values.
groupedBoxes = group_by(Boxes, Width) grsumBoxes=summarize(groupedBoxes, AvgLength = mean(Length)) grsumBoxes
## # A tibble: 7 × 2 ## Width AvgLength ## <dbl> <dbl> ## 1 3 8 ## 2 4 10 ## 3 5 20 ## 4 6 20 ## 5 7 11 ## 6 9 11 ## 7 10 3
Using the altered dataframe created in the previous slide, the following plot was created: 
This final example showcases every function that has been featured in this presentation. The plot of the final result is provided on the next slide.
summaryBoxes = Boxes %>%
mutate(Area = Length*Width*Height) %>%
select(Width, Length, Area) %>%
filter(Width <= 6) %>%
group_by(Width) %>%
summarize(AvgWidth = mean(Width),
AvgLength = mean(Length), AvgArea = mean(Area))
## # A tibble: 4 × 4 ## Width AvgWidth AvgLength AvgArea ## <dbl> <dbl> <dbl> <dbl> ## 1 3 3 8 480 ## 2 4 4 10 800 ## 3 5 5 20 2200 ## 4 6 6 20 2658