append_dest_origin: Reshape data into origin-destination pairs

In james-e-thomas/origindest: Reshape data into origin-destination pairs

Description

Requires dplyr and tidyr packages from tidyverse to be installed, origin_dest1 takes as its only argument x a dataframe (presumably containing information about locations) and outputs a dataframe containing all possible origin-destination pairings between the observations in the dataframe. A typical usage of this function would be to get a dataframe prepared for the subsequent calculation of distances or drive times between locations.

origin_dest2 instead takes two arguments/dataframes, one containing the origins (x) and the other containing the destinations (y).

origin_dest1 outputs a dataframe with:

• ncol(x)*2 columns;

• (nrow(x))*(nrow(x)-1) rows.

There are two times as many columns as in the original dataframe because each observation is now an origin-destination pair. This means that for each original column in the dataframe there is now an 'origin' column (whose name is appended with "origin_") and a 'destination' column (whose name is appended with "dest_").

There are (nrow(x))*(nrow(x)-1) rows because for each observation, there are nrow(x)-1 other locations for it to be paired with (we are not interested in the pairing of an observation with itself).

According to a similar logic, origin_dest2 will output a dataframe of dimensions (nrow(x)*nrow(y)) by (ncol(x) + ncol(y)). Note that if an observation appears on both the list of origins and destinations, the pairing of an observation with itself will have to be manually removed if it is not required.

Note that for origin_dest2, the choice or origin and destination matters. The list being produced is a list of possible combinations of A (origins) to B (destinations): the B to A combinations do not feature on the list produced. A possible usage would be to find pairings between customers and shops- the relevant shortest distance/drive time is that from the customer address (origin) to the shop address (destination). The shortest distance/drive time from the shop to the customer is unlikely to be of interest.

If the B to A combinations are of interest though, the append_dest_origin command can be used after origin_dest2. You may want to sort your observations following the use of this command.

Usage

origin_dest1(x)
origin_dest2(x, y)
append_dest_origin(x)

Arguments

x	dataframe to reshape; typically a list of locations. For origin_dest2 this is a list of origins. For append_dest_origin this should be the output of origin_dest2.
y	applicable to origin_dest2 only, this is a dataframe containing destination information.

Examples

# Example 1 -------------------------------------------------------------------

# Finding all pairs of US states

library(origindest)

state.name <- as.data.frame(state.name)

origin.dest.list <- origin_dest1(state.name)


# Example 2 -------------------------------------------------------------------

# Finding pairs of US states beginning with "A" and all other states not
# beginning with A

library(origindest)

not.beginning.with.A <- filter(as.data.frame(state.name),
                               state.name != "Alabama" &
                               state.name != "Alaska" &
                               state.name != "Arizona" &
                               state.name != "Arkansas")

beginning.with.A <- filter(as.data.frame(state.name),
                           state.name == "Alabama" |
                           state.name == "Alaska" |
                           state.name == "Arizona" |
                           state.name == "Arkansas")

origin.dest.list <- origin_dest2(not.beginning.with.A, beginning.with.A)

james-e-thomas/origindest documentation built on Aug. 20, 2020, 4:13 a.m.