Challenges

A suite of functions for:

• Data entry
• Data cleaning
• Accurate naming
• Visualization

Design Principles

• Open-source and open development
• Locally relevant
• Meet global standards
• Extensibility
• Simplicity

Installation

• Stable version:

install.packages("naijR")

• Development version (dev branch)

# install.packages('remotes')
remotes::install_github("BroVic/naijR", ref = "dev")

Administrative Regions: States

• Using strings i.e. character vectors

s <- c("Adamawa", "Bauchi", "Borno", "Gombe", "Taraba", "Yobe")
s

[1] "Adamawa" "Bauchi"  "Borno"   "Gombe"   "Taraba"  "Yobe"   

Administrative Regions: States

• Using states objects (S3 classes)

A special vector constructed with the states() function:

library(naijR)
states()

Abia

Adamawa

Akwa Ibom

Anambra

Bauchi

Bayelsa

Benue

Borno

Cross River

Delta

Ebonyi

Edo

Ekiti

Enugu

Federal Capital Territory

Gombe

Imo

Jigawa

Kaduna

Kano

Katsina

Kebbi

Kogi

Kwara

Lagos

Nasarawa

Niger

Ogun

Ondo

Osun

Oyo

Plateau

Rivers

Sokoto

Taraba

Yobe

Zamfara

Administrative Regions: States

# Using earlier created vector
(stateobj <- states(s))

Adamawa

Bauchi

Borno

Gombe

Taraba

Yobe

Administrative Regions: States

Objects representing the sub-national divisions inherit from an abstract class regions to confer a particular behaviour.

• regions is an abstract class i.e. it does not have constructible objects, but exists to define shared behaviour between states and lgas.

class(stateobj)

[1] "states"    "regions"   "character"

Administrative Regions: States

states has additional arguments:

function (states, gpz = NULL, all = TRUE, warn = TRUE) 
NULL

• gpz - a geopolitical zone (string)
• all - include FCT in the result? (logical)
• warn - notify if an element is not a valid State (logical)

Administrative Regions: Local Government Areas

• As with States, we can use character vectors with the names of the LGA.
• We can also create lgas objects - safer
• LGAs present an additional challenge:
  • Sheer number (774)
  • Duplication of LGAs
  • Ambiguity due to name-sharing

Because of this the function signature is more involved:

function (region = NA_character_, strict = FALSE, warn = TRUE) 
NULL

Note:

• region - i.e. one or more States (character vector only) or selected LGAs.
• strict - use LGA when there is name-sharing (logical, default is FALSE).
• warn - notify of wrong spelling (logical).

Helper Functions

• is_* - are elements of the object what they claim to be?
• fix_* - carry out repairs.

Example:

nas <- "Nassarawa"
is_state(nas)

[1] FALSE

nas <- fix_region(nas)

Error: Incorrect region name(s); consider reconstructing 'x' with `states()` or `lgas()` for a more reliable fix

nas <- fix_region(states(nas))
is_state(nas)

[1] TRUE

Fixing LGA spellings

am <-
  c("Amuwo-Olofin",
    "Amuwo-Odofin",
    "Amuwo-Odofin",
    "Amuwu-Odofin")

is_lga(am)

[1] FALSE  TRUE  TRUE FALSE

am |>
  fix_region() |>
  is_lga()

[1] TRUE TRUE TRUE TRUE

• Sometimes, LGAs cannot be repaired automatically
• This occurs when there are too name clashes
• The fixes can now be done interactively with the function fix_region_manual().
• See the article that describes how this is done by running the following code:

vignette("interactive", "naijR")

Phone Numbers

Deal with poorly entered phone numbers and MS Excel mutilations using fix_mobile.

Phone Numbers

• Input numeric values…

fix_mobile(8034510441)

[1] "08034510441"

Phone Numbers

• or strings…

fix_mobile("8034510441")

[1] "08034510441"

Numbers that cannot be repaired are turned into missing values i.e. NAs.

nums <- c("8034510441", "070304", "09O14358956")
fix_mobile(nums)

[1] "08034510441" NA            "09014358956"

Note that one of the digits of nums[3] is not 0 but O. The function automatically repairs it.

Maps

• Plain plots - by default shows State boundaries

map_ng()

Maps

map_ng(lgas())

args(map_ng)

function (region = character(), data = NULL, x = NULL, y = NULL, 
    breaks = NULL, categories = NULL, excluded = NULL, exclude.fill = NULL, 
    title = NULL, caption = NULL, show.neighbours = FALSE, show.text = FALSE, 
    legend.text = NULL, leg.x = deprecated(), leg.y = deprecated(), 
    leg.title, leg.orient = deprecated(), ...) 
NULL

• Input options
  • A collection of States or LGAs
  • A data frame
  • A collection of coordinates

Combining Concepts

What do you expect to be the result of the following code?

map_ng("Bauchi")

Consider the following possibilities:

• Bauchi is the name of a State in Nigeria.
• Bauchi is the name of an LGA in Bauchi State of Nigeria.
• We could draw a map of:
  • Bauchi State
  • All the LGAs in Bauchi State
  • Bauchi LGA
• This informed the polymorphism used in the package.

We can also create choropleth maps using the map_ng() function. For more info, read the vignette

vignette('nigeria-maps', 'naijR')

Some New Ideas

The package is not yet feature complete. Many changes still ahead:

• Provision of richer objects/methods
• Introduction of compiled code i.e. low-level constructs (C/C++)
• Link to Other Ecosystems
• A case for political wards
• More robust handling of phone numbers: Map to (inter)national standard
• Connection to geospatial packages

Collaboration

• The package is hosted publicly on GitHub and has a GPL-3 license, and thus open to modification, distribution, etc.
• How to contribute:
  • Issues
  • Pull_Requests
  • Documentation

Resources

• naijR website - http://brovic.github.io/naijR
• My personal blog - https://victorordu.wordpress.com

To contact me, visit my GitHub profile: https://github.com/BroVic