,
Mark Padgham [aut]| Title: | A Set of Tools For Solving The Maximal Covering Location Problem |
|---|---|
| Description: | Solving the "maximal covering location problem" as described by Church can be difficult for users not familiar with linear programming. maxcovr provides functions to make it easy to solve this problem, and tools to calculate facility coverage. |
| Authors: | Nicholas Tierney [aut, cre] ,
Mark Padgham [aut] |
| Maintainer: | Nicholas Tierney <[email protected]> |
| License: | GPL-3 |
| Version: | 0.1.3.9200 |
| Built: | 2024-09-24 04:14:48 UTC |
| Source: | https://github.com/njtierney/maxcovr |
This takes the proposed sites and the existing sites, with additional information from the model, and then returns a dataframe of all of the existing facilities that were relocated, and provides the distance to the nearest facility, which is presumably the location to which it was relocated to.
augment_facility_relocated(proposed_facility, existing_facility)augment_facility_relocated(proposed_facility, existing_facility)
proposed_facility |
facilities proposed for the model - but this data
has extra information ( |
existing_facility |
facilities existing for the model - but this data
has extra information ( |
dataframe
## Not run: mc_cv_n100_test %>% mutate(facility_distances = map2( .x = proposed_facility, .y = existing_facility, .f = augment_facility_relocated)) %>% select(facility_distances) %>% .[[1]] ## End(Not run)## Not run: mc_cv_n100_test %>% mutate(facility_distances = map2( .x = proposed_facility, .y = existing_facility, .f = augment_facility_relocated)) %>% select(facility_distances) %>% .[[1]] ## End(Not run)
This returns the user dataframe, with added columns containing distance
between that user and a given facility - IDs are generated for IDs and
facilities that correspond to their row number.
augment_user(facilities_selected, existing_facilities, existing_users)augment_user(facilities_selected, existing_facilities, existing_users)
facilities_selected |
dataframe of facilities selected, obtained from
|
existing_facilities |
existing facilities |
existing_users |
existing users |
tibble of users, with distances between each user and facility
## Not run: mc_facilities_selected <-extract_facility_selected( solution_vector = x$lp_solution$solution, A_mat = x$A, proposed_facilities = x$proposed_facility) augmented_users <- augment_user( facilities_selected = mc_facilities_selected, existing_facilities = mc_cv_fit_n20_test_1$existing_facility, existing_users = mc_cv_fit_n20_test_1$existing_user ) ## End(Not run)## Not run: mc_facilities_selected <-extract_facility_selected( solution_vector = x$lp_solution$solution, A_mat = x$A, proposed_facilities = x$proposed_facility) augmented_users <- augment_user( facilities_selected = mc_facilities_selected, existing_facilities = mc_cv_fit_n20_test_1$existing_facility, existing_users = mc_cv_fit_n20_test_1$existing_user ) ## End(Not run)
This function is wrapper to nearest, adding is_covered to the model. This
function is explicit about inputs, and is useful in cross validation -
evaluating how test data performs against suggested facilities in the
training set. This might be added to nearest, and may become obsolete.
augment_user_tested(all_facilities, test_data, distance_threshold = 100)augment_user_tested(all_facilities, test_data, distance_threshold = 100)
all_facilities |
data.frame Facilities selected in maxcovr model |
test_data |
data.frame test data (but it could be any |
distance_threshold |
numeric |
dataframe containing distances between each test data observation and the nearest facility.
## Not run: mc_cv_relocate_n100_cut %>% mutate(user_nearest_test = map2( .x = facilities_selected, .y = test, .f = augment_user_tested )) ## End(Not run)## Not run: mc_cv_relocate_n100_cut %>% mutate(user_nearest_test = map2( .x = facilities_selected, .y = test, .f = augment_user_tested )) ## End(Not run)
This is a wrapper function that returns a logical matrix, of 1 if distance between element i, j is less than or equal to the distance_cutoff, and 0 otherwise.
binary_distance_matrix(facility, user, distance_cutoff, d_proposed_user = NULL)binary_distance_matrix(facility, user, distance_cutoff, d_proposed_user = NULL)
facility |
data.frame of facilities |
user |
data.frame of users |
distance_cutoff |
integer of distance to use for cutoff |
d_proposed_user |
Option distance matrix between proposed facilities and users (see Examples). |
a logical matrix, of 1 if distance between element i, j is less than or equal to the distance_cutoff, and 0 otherwise.
Create a binary matrix TRUE if distance satisfies a condition
binary_matrix_cpp(facility, user, distance_cutoff)binary_matrix_cpp(facility, user, distance_cutoff)
facility |
a matrix with longitude and latitude in the first two columns |
user |
a matrix with longitude and latitude in the first two columns |
distance_cutoff |
numeric indicating the distance cutoff (in metres) you are interested in. If a number is less than distance_cutoff, it will be 1, if it is greater than it, it will be 0. |
a logical matrix 1 if distance between element i, j is less than or equal to the distance_cutoff, and 0 otherwise
In the york building and york crime context, writing
nearest(york_crime,york) reads as "find the nearest crime in york to
each building in york, and returns a dataframe with every building in
york, the nearest york_crime to each building, and the distance in
metres between the two."
coverage(nearest_df, to_df, distance_cutoff = 100, ...)coverage(nearest_df, to_df, distance_cutoff = 100, ...)
nearest_df |
dataframe containing latitude and longitude |
to_df |
dataframe containing latitude and longitude |
distance_cutoff |
integer the distance threshold you are interested in assessing coverage at |
... |
extra arguments to pass to nearest |
a dataframe containing information about the distance threshold uses (distance_within), the number of events covered and not covered (n_cov, n_not_cov), the percentage covered and not covered (pct_cov, pct_not_cov), and the average distance and sd distance.
library(dplyr) # already existing locations york_selected <- york %>% filter(grade == "I") # proposed locations york_unselected <- york %>% filter(grade != "I") coverage(york_selected, york_crime) coverage(york_crime, york_selected)library(dplyr) # already existing locations york_selected <- york %>% filter(grade == "I") # proposed locations york_unselected <- york %>% filter(grade != "I") coverage(york_selected, york_crime) coverage(york_crime, york_selected)
Convert from degrees to radians
deg2rad(deg)deg2rad(deg)
deg |
A numeric vector in units of degrees. |
The input numeric vector, converted to units of radians.
Convert degrees to radians
deg2rad_cpp(deg)deg2rad_cpp(deg)
deg |
degrees |
radians
Create a matrix of distances between two areas
distance_matrix_cpp(facility, user)distance_matrix_cpp(facility, user)
facility |
a matrix with longitude and latitude in the first two columns |
user |
a matrix with longitude and latitude in the first two columns |
a matrix of distances in metres between each user and facility, with nrow(user) rows and nrow(facility) columns.
Extract the existing coverage
extract_existing_coverage(existing_facilities, existing_users, distance_cutoff)extract_existing_coverage(existing_facilities, existing_users, distance_cutoff)
existing_facilities |
the existing facilities |
existing_users |
the existing users |
distance_cutoff |
the distance cutoffs |
tibble of existing coverage
## Not run: extract_existing_coverage(existing_facilities = x$existing_facility, existing_users = x$existing_user, distance_cutoff = x$distance_cutoff) ## End(Not run)## Not run: extract_existing_coverage(existing_facilities = x$existing_facility, existing_users = x$existing_user, distance_cutoff = x$distance_cutoff) ## End(Not run)
This takes the linear programming solution, the A matrix, and the proposed facilities. It returns a tibble, which contains the facilities chosen from the proposed facilities.
extract_facility_selected(solution_vector, A_mat, proposed_facilities)extract_facility_selected(solution_vector, A_mat, proposed_facilities)
solution_vector |
vector from lp_solution$solution |
A_mat |
The "A" matrix from the solver |
proposed_facilities |
Dataframe of proposed facilities |
dataframe of selected facilities
# assuming that you've run max_coverage using lpSolve, then you # will save the model output before the extraction process # as `x`. ## Not run: mc_facilities_selected <- extract_facility_selected( solution_vector = x$lp_solution$solution, A_mat = x$A, proposed_facilities = x$proposed_facility) ## End(Not run)# assuming that you've run max_coverage using lpSolve, then you # will save the model output before the extraction process # as `x`. ## Not run: mc_facilities_selected <- extract_facility_selected( solution_vector = x$lp_solution$solution, A_mat = x$A, proposed_facilities = x$proposed_facility) ## End(Not run)
extract_mc_results takes a fitted max_coverage object and
returns useful summary information from the model. It exists so that the
manipulation functions for the outcomes from the solver have another
home - this makes it easier to maintain this package, and heeds to this
idea of having functions that are specialised. The name of this function
is likely to change in the near future.
extract_mc_results(x)extract_mc_results(x)
x |
the fitted model from |
a list containing multiple dataframes summarising the model
extract_mc_results_relocation takes a fitted max_coverage object and
returns useful summary information from the model, specifically for the
relocation method.
extract_mc_results_relocation(x)extract_mc_results_relocation(x)
x |
the fitted model from max_coverage_relocation |
a list containing multiple dataframes summarising the model
This function takes the users information, the distance cutoff, and the number of facilities added, and then returns a one-row dataframe containing summary information about the coverage.
extract_model_coverage(augmented_user, distance_cutoff, n_added)extract_model_coverage(augmented_user, distance_cutoff, n_added)
augmented_user |
dataframe obtained from |
distance_cutoff |
numeric of the distance cutoff |
n_added |
numeric of the number of facilities added |
tibble of summary coverage info
## Not run: augmented_users <- augment_user( facilities_selected = mc_facilities_selected, existing_facilities = x$existing_facility, existing_users = x$existing_user) extract_model_coverage( augmented_user = augmented_users, distance_cutoff = x$distance_cutoff, n_added = x$n_added) ## End(Not run)## Not run: augmented_users <- augment_user( facilities_selected = mc_facilities_selected, existing_facilities = x$existing_facility, existing_users = x$existing_user) extract_model_coverage( augmented_user = augmented_users, distance_cutoff = x$distance_cutoff, n_added = x$n_added) ## End(Not run)
Extract additional users affected by new coverage from the new facilities
extract_users_affected(A_mat, solution_vector, user_id, users_not_covered)extract_users_affected(A_mat, solution_vector, user_id, users_not_covered)
A_mat |
A matrix |
solution_vector |
The vector of solutions |
user_id |
The IDs of the individuals |
users_not_covered |
those users not covered by original AEDs |
tibble taken from users, those who are affectd by new placements
## Not run: extract_users_affected( A_mat = x$A, solution_vector = x$lp_solution$solution, user_id = x$user_id, users_not_covered = x$user_not_covered) ## End(Not run)## Not run: extract_users_affected( A_mat = x$A, solution_vector = x$lp_solution$solution, user_id = x$user_id, users_not_covered = x$user_not_covered) ## End(Not run)
Uses haversines formula to calculate the distance between lat/long co-ordinates of every facility and every user, returning a data_frame. You can think of "facilities" as something like mobile towers, police centres, or AED locations, and "users" as something like individual houses, crime locations, or heart attack locations. The motivating example for this function was finding the distance from Automatic Electronic Defibrillators (AEDs) to each Out of Hospital Cardiac Arrest (OHCA), where the locations for AEDs and OHCAs are in separate dataframes. Currently facifacility_user_dist makes the strict assumption that the facility and user dataframes have columns named aed_id, lat, and long, and ohca_id, lat, and long. This will be updated soon.
facility_user_dist(facility, user, coverage_distance = 100, nearest = "facility")facility_user_dist(facility, user, coverage_distance = 100, nearest = "facility")
facility |
a dataframe containing columns named "lat", and "long". |
user |
a dataframe containing columns "lat", and "long". |
coverage_distance |
numeric indicating the coverage level for the facilities to be within in metres to a user. Default value is 100 metres. |
nearest |
character Can be "facility", "user", and "both". Defaults to "facility". When set to "facility", returns a dataframe where every row refers to every user, and the closest facility to each user. When set to "user", it returns a dataframe where every row is every facility, and the closest user to each facility. When set to "both", which will return every pairwise combination of distances. Be careful when default is "facility". |
a data frame containing the two datasets joined together with columns named facility_id, lat_facility, long_facility, user_id, lat_user, long_user, distance in meters between each the given facility and user in a row.
This is a data manipulation function for facility_user_dist. This function creates a spread matrix of the distances between each ohca and each aed. There is an ohca_id column, and then a column for each aed_id, with a given cell being the distance between an ohca in a row, and that column. This distance is converted into an indicator variable, based upon whether that distance is less than the provided dist_indic parameter. In the future I might change the dist_indic function to be optional, but this whole function mainly exists to make it easier to do the computation in the max_coverage function.
facility_user_indic(df_dist, dist_indic)facility_user_indic(df_dist, dist_indic)
df_dist |
dataframe from facility_user_dist. Requires nearest = "both" |
dist_indic |
an indicator of the distance you want to be TRUE / FALSE |
dataframe with variables ohca_id, and aed_id_number, with the id from each aed_id being transposed into each column name.
(Internal) Create a dataframe of the users not covered
find_users_not_covered(existing_facility, user, distance_cutoff, d_existing_user = NULL)find_users_not_covered(existing_facility, user, distance_cutoff, d_existing_user = NULL)
existing_facility |
data.frame of existing facilities |
user |
data.frame of existing users |
distance_cutoff |
integer of distance cutoff |
d_existing_user |
Optional distance matrix between existing facilities and users. |
data.frame of those users not covered by current facilities
Test if the object is a maxcovr object
is.maxcovr(x)is.maxcovr(x)
x |
An object |
TRUE if the object inherits from the maxcovr class.
Test if the object is a maxcovr_relocation object
is.maxcovr_relocation(x)is.maxcovr_relocation(x)
x |
An object |
TRUE if the object inherits from the maxcovr_relocation class.
max_coverage solves the binary optimisation problem known as the
"maximal covering location problem" as described by Church
(http://www.geo .ucsb.edu/~forest/G294download/MAX_COVER_RLC_CSR.pdf).
This package was implemented to make it easier to solve this problem in the
context of the research initially presented by Chan et al
(http://circ.ahajournals.org/content/127/17/1801.short) to identify ideal
locations to place AEDs.
max_coverage(existing_facility, proposed_facility, user, distance_cutoff, n_added, d_existing_user = NULL, d_proposed_user = NULL, solver = "glpk")max_coverage(existing_facility, proposed_facility, user, distance_cutoff, n_added, d_existing_user = NULL, d_proposed_user = NULL, solver = "glpk")
existing_facility |
data.frame containing the facilities that are already in existing, with columns names lat, and long. |
proposed_facility |
data.frame containing the facilities that are being proposed, with column names lat, and long. |
user |
data.frame containing the users of the facilities, along with column names lat, and long. |
distance_cutoff |
numeric indicating the distance cutoff (in metres) you are interested in. If a number is less than distance_cutoff, it will be 1, if it is greater than it, it will be 0. |
n_added |
the maximum number of facilities to add. |
d_existing_user |
Optional distance matrix between existing facilities and users. Default distances are direct (geospherical ellipsoidal) distances; this allows alternative measures such as street-network distances to be submitted (see Examples). |
d_proposed_user |
Option distance matrix between proposed facilities and users (see Examples). |
solver |
character "glpk" (default) or "lpSolve". "gurobi" is currently in development, see https://github.com/njtierney/maxcovr/issues/25 |
dataframe of results
library(dplyr) # already existing locations york_selected <- york %>% filter(grade == "I") # proposed locations york_unselected <- york %>% filter(grade != "I") mc_result <- max_coverage(existing_facility = york_selected, proposed_facility = york_unselected, user = york_crime, distance_cutoff = 100, n_added = 20) mc_result summary(mc_result) # get the facilities chosen mc_result$facility_selected # get the users affected mc_result$user_affected # get the summaries mc_result$summary # Example of street-network distance calculations ## Not run: library(dodgr) net <- dodgr_streetnet_sf ("york england") %>% weight_streetnet (wt_profile = "foot") from <- match_points_to_graph (v, york_selected [, c ("long", "lat")]) to <- match_points_to_graph (v, york_crime [, c ("long", "lat")]) d_existing_user <- dodgr_dists (net, from = from, to = to) from <- match_points_to_graph (v, york_unselected [, c ("long", "lat")]) d_proposed_user <- dodgr_dists (net, from = from, to = to) mc_result <- max_coverage(existing_facility = york_selected, proposed_facility = york_unselected, user = york_crime, distance_cutoff = 100, n_added = 20, d_existing_user = d_existing_user, d_proposed_user = d_proposed_user) ## End(Not run)library(dplyr) # already existing locations york_selected <- york %>% filter(grade == "I") # proposed locations york_unselected <- york %>% filter(grade != "I") mc_result <- max_coverage(existing_facility = york_selected, proposed_facility = york_unselected, user = york_crime, distance_cutoff = 100, n_added = 20) mc_result summary(mc_result) # get the facilities chosen mc_result$facility_selected # get the users affected mc_result$user_affected # get the summaries mc_result$summary # Example of street-network distance calculations ## Not run: library(dodgr) net <- dodgr_streetnet_sf ("york england") %>% weight_streetnet (wt_profile = "foot") from <- match_points_to_graph (v, york_selected [, c ("long", "lat")]) to <- match_points_to_graph (v, york_crime [, c ("long", "lat")]) d_existing_user <- dodgr_dists (net, from = from, to = to) from <- match_points_to_graph (v, york_unselected [, c ("long", "lat")]) d_proposed_user <- dodgr_dists (net, from = from, to = to) mc_result <- max_coverage(existing_facility = york_selected, proposed_facility = york_unselected, user = york_crime, distance_cutoff = 100, n_added = 20, d_existing_user = d_existing_user, d_proposed_user = d_proposed_user) ## End(Not run)
This function adds a relocation step
max_coverage_relocation(existing_facility = NULL, proposed_facility, user, distance_cutoff, cost_install, cost_removal, cost_total, solver = "lpSolve", return_early = FALSE)max_coverage_relocation(existing_facility = NULL, proposed_facility, user, distance_cutoff, cost_install, cost_removal, cost_total, solver = "lpSolve", return_early = FALSE)
existing_facility |
data.frame containing the facilities that are already in existing, with columns names lat, and long. |
proposed_facility |
data.frame containing the facilities that are being proposed, with column names lat, and long. |
user |
data.frame containing the users of the facilities, along with column names lat, and long. |
distance_cutoff |
numeric indicating the distance cutoff (in metres) you are interested in. If a number is less than distance_cutoff, it will be 1, if it is greater than it, it will be 0. |
cost_install |
integer the cost of installing a new facility |
cost_removal |
integer the cost of removing a facility |
cost_total |
integer the total cost allocated to the project |
solver |
character "glpk" (default) or "lpSolve". "gurobi" is currently in development, see https://github.com/njtierney/maxcovr/issues/25 |
return_early |
logical TRUE if I do not want to run the extraction process, FALSE if I want to just return the lpsolve model etc. |
dataframe of results
## Not run: library(dplyr) # subset to be the places with towers built on them. york_selected <- york %>% filter(grade == "I") york_unselected <- york %>% filter(grade != "I") # OK, what if I just use some really crazy small data to optimise over. # mc_relocate <- max_coverage_relocation(existing_facility = york_selected, proposed_facility = york_unselected, user = york_crime, distance_cutoff = 100, cost_install = 5000, cost_removal = 200, cost_total = 600000) mc_relocate summary(mc_relocate) ## End(Not run)## Not run: library(dplyr) # subset to be the places with towers built on them. york_selected <- york %>% filter(grade == "I") york_unselected <- york %>% filter(grade != "I") # OK, what if I just use some really crazy small data to optimise over. # mc_relocate <- max_coverage_relocation(existing_facility = york_selected, proposed_facility = york_unselected, user = york_crime, distance_cutoff = 100, cost_install = 5000, cost_removal = 200, cost_total = 600000) mc_relocate summary(mc_relocate) ## End(Not run)
Using the model-modified dataframe of proposed_facility, count the
number of events installed.
n_installed(proposed_facility)n_installed(proposed_facility)
proposed_facility |
dataframe from the mc_model, of facilities proposed
with the additional information about whether the facility was installed
or not - |
datafrmae
## Not run: mc_cv_n100_test %>% mutate(n_installed = map( .x = proposed_facility, .f = n_installed )) %>% select(n_installed) %>% .[[1]] ## End(Not run)## Not run: mc_cv_n100_test %>% mutate(n_installed = map( .x = proposed_facility, .f = n_installed )) %>% select(n_installed) %>% .[[1]] ## End(Not run)
Extract the number of facilities relocated.
n_relocated(existing_facility)n_relocated(existing_facility)
existing_facility |
the facilities originally existing, as output from
the model (e.g., |
dataframe containing one column of the number of things relocated
## Not run: mc_cv_n100_test %>% mutate(n_relocated = map( .x = existing_facility, .f = n_relocated)) %>% select(n_relocated) %>% .[[1]] ## End(Not run)## Not run: mc_cv_n100_test %>% mutate(n_relocated = map( .x = existing_facility, .f = n_relocated)) %>% select(n_relocated) %>% .[[1]] ## End(Not run)
This function finds the nearest lat/long pairs to another lat/long pair.
So in the york building and york crime context, writing
nearest(york_crime,york) reads as "find the nearest crime in york to
each building in york, and returns a dataframe with every building in york,
the nearest york_crime to each building, and the distance in metres between
the two. Likewise, you could write nearest(york, york_crime), and this
would return the nearest building to every crime. nearest assumes that
the names of the latitude and longitude are "lat" and "long", but you can
provide these names.
nearest(nearest_df, to_df, nearest_lat = "lat", nearest_long = "long", to_lat = "lat", to_long = "long")nearest(nearest_df, to_df, nearest_lat = "lat", nearest_long = "long", to_lat = "lat", to_long = "long")
nearest_df |
a dataframe containing latitude and longitude |
to_df |
a dataframe containing latitude and longitude |
nearest_lat |
name of latitude in nearest_df |
nearest_long |
name of longitude in nearest_df |
to_lat |
name of latitude in to_df |
to_long |
name of longitude in to_df |
dataframe of "to_df" along with the nearest "nearest_df" to each row, along with the distance between the two, and the nearest_id, the row position of the nearest_df closest to that row.
library(maxcovr) nearest(nearest_df = york_crime, to_df = york) # you can use the pipe as well ## Not run: library(magrittr) york_crime %>% nearest(york) ## End(Not run)library(maxcovr) nearest(nearest_df = york_crime, to_df = york) # you can use the pipe as well ## Not run: library(magrittr) york_crime %>% nearest(york) ## End(Not run)
nearest facility + distance to a user
nearest_facility_dist(facility, user)nearest_facility_dist(facility, user)
facility |
a matrix with longitude and latitude in the first two columns |
user |
a matrix with longitude and latitude in the first two columns |
matrix with 3 columns: user_id, facility_id, distance, where the user_id is the identifier for the user, the facility_id is the identifier for the facility that is closest to that user, and the distance is the distance in metres from that user to that facility.
This function is a wrapper for the similarly named, nearest_facility_dist
function used inside max_coverage to calculate distances
so that the nearest facilities can be found.
nearest_facility_distances(existing_facility, user)nearest_facility_distances(existing_facility, user)
existing_facility |
dataframe of existing facilities |
user |
dataframe of users to place facilities to cover |
A tibble with 3 columns: user_id, facility_id, distance, where the user_id is the identifier for the user, the facility_id is the identifier for the facility that is closest to that user, and the distance is the distance in metres from that user to that facility.
This function uses the haversine formula to calculate the great circle distance between two locations, identified by their latitudes and longitudes. It is borrowed from rnoaa (https://github.com/ropenscilabs/rnoaa/blob/master/R/meteo_distance.R) and included here as rnoaa is a large package that is rather unrelated to maxcovr. I have renamed it from meteo_spherical_distance to spherical_distance
spherical_distance(lat1, long1, lat2, long2)spherical_distance(lat1, long1, lat2, long2)
lat1 |
Latitude of the first location. |
long1 |
Longitude of the first location. |
lat2 |
Latitude of the second location. |
long2 |
Longitude of the second location. |
A numeric value giving the distance in meters between the pair of locations.
This function assumes an earth radius of 6,371 km.
Alex Simmons [email protected], Brooke Anderson [email protected]
spherical_distance(lat1 = -27.4667, long1 = 153.0217, lat2 = -27.4710, long2 = 153.0234)spherical_distance(lat1 = -27.4667, long1 = 153.0217, lat2 = -27.4710, long2 = 153.0234)
Calculate distance using haversines formula
spherical_distance_cpp(lat1, long1, lat2, long2)spherical_distance_cpp(lat1, long1, lat2, long2)
lat1 |
latitude from the first location |
long1 |
longitude from the first location |
lat2 |
latitude from the second location |
long2 |
longitude from the second location |
distance in metres between two locations
Calculate (vectorized) distance using haversines formula
spherical_distance_cpp_vec(lat1, long1, lat2, long2)spherical_distance_cpp_vec(lat1, long1, lat2, long2)
lat1 |
latitude from the first location |
long1 |
longitude from the first location |
lat2 |
latitude from the second location |
long2 |
longitude from the second location |
distance in metres between two locations
Provides summary information of the coverage, using the distance dataframe
created by facility_user_dist().
summarise_coverage(df_dist, distance_cutoff = 100)summarise_coverage(df_dist, distance_cutoff = 100)
df_dist |
distance matrix, as computed by facility_user_dist |
distance_cutoff |
the critical distance range that you would like to know. The default is 100m. |
dataframe
This takes data from the function augment_facility_relocated function of
the same name and then summarises it to find the average and sd of the
distance between the two.
summarise_relocated_dist(augment_facility_relocated)summarise_relocated_dist(augment_facility_relocated)
augment_facility_relocated |
dataframe from function:
|
dataframe
## Not run: mc_cv_n100_test %>% mutate( facility_distances = map2( .x = proposed_facility, .y = existing_facility, .f = augment_facility_relocated ), summary_relocated_dist = map( .x = facility_distances, .f = summarise_relocated_dist ) ) %>% # select(facility_distances) %>% select(summary_relocated_dist) %>% .[[1]] ## End(Not run)## Not run: mc_cv_n100_test %>% mutate( facility_distances = map2( .x = proposed_facility, .y = existing_facility, .f = augment_facility_relocated ), summary_relocated_dist = map( .x = facility_distances, .f = summarise_relocated_dist ) ) %>% # select(facility_distances) %>% select(summary_relocated_dist) %>% .[[1]] ## End(Not run)
This uses a user dataframe obtained from something like
augment_user_tested.
summarise_user_cov(user)summarise_user_cov(user)
user |
dataframe of users with distances between each user and the
nearest facility ( |
dataframe containing information on the number of users, the number of events covered, the proportion of events covered, and the distance from each
## Not run: summarise_user_cov(augmented_user_test) ## End(Not run)## Not run: summarise_user_cov(augmented_user_test) ## End(Not run)
Summary for max_coverage cross validation
summary_mc_cv(model, test_data)summary_mc_cv(model, test_data)
model |
the cross validated model |
test_data |
the cross validated test data |
a summary dataframe
## Not run: library(maxcovr) library(tidyverse) york_selected <- york %>% filter(grade == "I") york_unselected <- york %>% filter(grade != "I") mc_cv_fixed <- modelr::crossv_kfold(york_crime, 5) %>% mutate(test = map(test,as_tibble), train = map(train,as_tibble)) mc_cv_fit <- map_df(mc_cv_fixed$train, ~max_coverage(existing_facility = york_selected, proposed_facility = york_unselected, user = ., n_added = 20, distance_cutoff = 100)) summary_mc_cv(mc_cv_fit, mc_cv_fixed$test) ## End(Not run)## Not run: library(maxcovr) library(tidyverse) york_selected <- york %>% filter(grade == "I") york_unselected <- york %>% filter(grade != "I") mc_cv_fixed <- modelr::crossv_kfold(york_crime, 5) %>% mutate(test = map(test,as_tibble), train = map(train,as_tibble)) mc_cv_fit <- map_df(mc_cv_fixed$train, ~max_coverage(existing_facility = york_selected, proposed_facility = york_unselected, user = ., n_added = 20, distance_cutoff = 100)) summary_mc_cv(mc_cv_fit, mc_cv_fixed$test) ## End(Not run)
Summary for max_coverage cross validation for relocation models
summary_mc_cv_relocate(model, test_data)summary_mc_cv_relocate(model, test_data)
model |
the cross validated model |
test_data |
the cross validated test data |
a summary dataframe
## Not run: library(maxcovr) library(tidyverse) york_selected <- york %>% filter(grade == "I") york_unselected <- york %>% filter(grade != "I") mc_cv <- modelr::crossv_kfold(york_crime, 5) %>% mutate(test = map(test,as_tibble), train = map(train,as_tibble)) mc_cv_relocate <- map_df(mc_cv$train, ~max_coverage_relocation(existing_facility = york_selected, proposed_facility = york_unselected, user = ., cost_install = 2500, cost_removal = 700, cost_total = 50000, distance_cutoff = 100, solver = "gurobi")) summary_mc_cv_relocate(mc_cv_relocate, mc_cv$test) ## End(Not run)## Not run: library(maxcovr) library(tidyverse) york_selected <- york %>% filter(grade == "I") york_unselected <- york %>% filter(grade != "I") mc_cv <- modelr::crossv_kfold(york_crime, 5) %>% mutate(test = map(test,as_tibble), train = map(train,as_tibble)) mc_cv_relocate <- map_df(mc_cv$train, ~max_coverage_relocation(existing_facility = york_selected, proposed_facility = york_unselected, user = ., cost_install = 2500, cost_removal = 700, cost_total = 50000, distance_cutoff = 100, solver = "gurobi")) summary_mc_cv_relocate(mc_cv_relocate, mc_cv$test) ## End(Not run)
Listed buildings provided by the City of York Council, made available here: https://data.gov.uk/dataset/listed-buildings24/resource/8c32fb55-0e40-457f-98f9-6494503e283b. This data contains public sector information licensed under the Open Government Licence v3.0: https://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/.
yorkyork
A data frame with seven variables: long, lat,
object_id, desig_id, pref_ref, name, and
grade.
longlongitude of the building
latlatitude of the building
object_idunique identifier for the building
desig_idID related to a feature that is not yet known to me
pref_refID related to a feature that is not yet known to me
namename of the building
gradeone of the three (I, II, III) cateogories of listed buildings
For further details, see https://www.york.gov.uk/info/20215/conservation_and_listed_buildings/1346/listed_buildings and https://data.gov.uk/dataset/listed-buildings24/resource/8c32fb55-0e40-457f-98f9-6494503e283b
Crime locations obtained via the ukpolice R package: https://github.com/njtierney/ukpolice, which uses the data made available through the uk crime API:<data.police.uk/docs/>. This data contains public sector information licensed under the Open Government Licence v3.0: https://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/.
york_crimeyork_crime
A data frame with variables: category, persistent_id,
date, lat, long, street_id, street_name, context, id, location_type, location_subtype, and outcome_status.
'
category: Category of the crime (https://data.police.uk/docs/method/crime-street/)
persistent_id: 64-character unique identifier for that crime. (This is different to the existing 'id' attribute, which is not guaranteed to always stay the same for each crime.)
date: Date of the crime YYYY-MM
latitude: Latitude
longitude: Longitude
street_id: Unique identifier for the street
street_name: Name of the location. This is only an approximation of where the crime happened
context: Extra information about the crime (if applicable)
id: ID of the crime. This ID only relates to the API, it is NOT a police identifier
location_type: The type of the location. Either Force or BTP: Force indicates a normal police force location; BTP indicates a British Transport Police location. BTP locations fall within normal police force boundaries.
location_subtype: For BTP locations, the type of location at which this crime was recorded.
outcome_status: The category and date of the latest recorded outcome for the crime
more documentation here: https://data.police.uk/docs/method/crime-street/
For further details, see https://www.york.gov.uk/info/20215/conservation_and_listed_buildings/1346/listed_buildings and https://data.gov.uk/dataset/listed-buildings24/resource/8c32fb55-0e40-457f-98f9-6494503e283b