EcoCounteR - An R API wrapper for the EcoCounter project

Installation

Until now the package is not on CRAN but you can download it via GitHub with the following command:

if (!require("devtools"))
  install.packages("devtools")
devtools::install_github("JohannesFriedrich/EcoCounteR")

Introduction

The R-package EcoCounteR is an API wrapper for data collected by https://www.eco-compteur.com. You can request available data from different Eco Counter installations by submitting the Id of the installation to different functions:

Function name	Description	Example
get_EcoCounter_info()	request common information about an installation	get_EcoCounter_info(100020207)
get_EcoCounter_location()	request location of an installation	get_EcoCounter_location(100020207)
get_EcoCounter_data()	request data from an installation	get_EcoCounter_data(101020207)

The following chapters show some examples how to use the R-package EcoCounteR.

Request basic information from Eco Counter API

First we load some needed packages. The function get_EcoCounter_info() will deliver some information about an installation located at Calgary in the Stephen Avenue. The only argument in get_EcoCounter_info() is the Id of the installation.

library(EcoCounteR)
library(leaflet)
library(htmltools)

EcoCounterId <- 100020207
info <- get_EcoCounter_info(EcoCounterId)

info

id	name	nameId
101020207	Stephen Avenue Pedestrian Counts	Z_101020207
102020207	Stephen Avenue Cyclist Counts	Z_102020207
103020207	Stephen Avenue Car Counts	Z_103020207

For later requests we need the coloumn id to get data from this specific installation.

The function get_EcoCounter_info() can also handle multiple EcoCounterIds, e.g.:

EcoCounterIds <- c(100020207, 100017181)

info <- get_EcoCounter_info(EcoCounterIds)

info

id	name	nameId
101020207	Stephen Avenue Pedestrian Counts	Z_101020207
102020207	Stephen Avenue Cyclist Counts	Z_102020207
103020207	Stephen Avenue Car Counts	Z_103020207
101017181	7 Street South of 3 Ave SW_SB	Z_IN_101017181
102017181	7 Street South of 3 Ave SW_NB	Z_OUT_102017181

Map location of an installation

To get an overview of the installations, we can use the function get_EcoCounter_location(). This function extracts the coordinates from the installations. We can insert the result directly to the R-package leaflet.

get_EcoCounter_location(EcoCounterId) %>% 
leaflet() %>%
  addProviderTiles(providers$OpenStreetMap) %>% 
  addTiles() %>%  
  addMarkers(~long, ~lat, popup = ~htmltools::htmlEscape(name))

Request data from Eco Counter API

Now we can use the extracted information to request data. First we load some packages which help to analyse and visualise data.

library(dplyr)
library(ggplot2)
library(scales)

From section 2 we know different sub-Ids of an installation. We allocate the different sub-Ids:

Stephan_Avenue_Ped <- 101020207
Stephan_Avenue_Bike <- 102020207
Stephan_Avenue_Car <- 103020207

Now it’s time to request some data. For that purpose the function get_EcoCounter_data() can be used. The function takes just one required argument: EcoCounterId. Note: This is not the same EcoCounterId as inserted in get_EcoCounter_data()and get_EcoCounter_location(). The function get_EcoCounter_data() can take more arguments to specify the request.

Argument	Description	Example
EcoCounterId	sub-Id to request specific installation	101020207
from	Start date of the requested data in the form “YYYYMMDD”	“20180101”
to	End date of the requested data in the form “YYYYMMDD”	“20180110”
step	Frequency of saved data: 2 (15 min), 3 (hourly), 4 (daily), 5 (weekly), 6 (monthly), 7 (yearly)	step = 3

Some examples:

## get daily saved data for the day before yesterday and yesterday
get_EcoCounter_data(101020207)

## get daily saved data from 1st January 2018 until yesterday 
get_EcoCounter_data(101020207, from = "20180101")

## get hourly saved data from 1st January 2018 to 31st January 2018 
get_EcoCounter_data(101020207, from = "20180101", to = "20180131", step = 3)

Now let’s put things into practical terms. We request data from 1st to 20th January 2018. With mutate(Type = XXX) we add a new coloum to the data.frame with the specific type of request.

Stephan_Avenue_Ped_daily <- get_EcoCounter_data(Stephan_Avenue_Ped, from = "20180101", to = "20180131") %>% 
  mutate(Type = "Pedestrains")

Stephan_Avenue_Bike_daily <- get_EcoCounter_data(Stephan_Avenue_Bike, from = "20180101", to = "20180131") %>% 
  mutate(Type = "Bikes")

Stephan_Avenue_Car_daily <- get_EcoCounter_data(Stephan_Avenue_Car, from = "20180101", to = "20180131") %>% 
  mutate(Type = "Cars")

Stephen_Avenue_complete <- bind_rows(Stephan_Avenue_Ped_daily, Stephan_Avenue_Bike_daily, Stephan_Avenue_Car_daily)
## Warning in bind_rows_(x, .id): Unequal factor levels: coercing to character
## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector

## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector

## Warning in bind_rows_(x, .id): binding character and factor vector,
## coercing into character vector

In the last line all single data.frames are combined to a single data.frame. You can further analyse this data to your own needs, e.g. find the days with the most appearance per type:

Stephen_Avenue_complete %>% 
  group_by(Type) %>% 
  top_n(1, comptage)
## # A tibble: 3 x 4
## # Groups:   Type [3]
##   date                comptage id        Type       
##   <dttm>                 <int> <chr>     <chr>      
## 1 2018-01-19 00:00:00    13436 101020207 Pedestrains
## 2 2018-01-19 00:00:00      532 102020207 Bikes      
## 3 2018-01-27 00:00:00     2349 103020207 Cars

Visualise Data

The following examples shows the visualisation of the data.

g <- ggplot(Stephen_Avenue_complete, aes(x = date, y = comptage, fill = Type)) +
  geom_bar(stat = "identity") +
  scale_x_datetime(labels = date_format("%d.%m.%Y")) +
  theme(legend.position = "bottom",
        axis.text.x = element_text(angle = 45, hjust = 1))

plot(g)

g2 <- g +
  facet_wrap(~Type, scales = "free_y") +
  geom_smooth(show.legend = FALSE)

plot(g2)

A more comprehensive visualisation

We load all available data with the best resolution (every 15 min: step = 2). Afterwars an animated heatmap is created with the R-package gganimate (Note: the synatx presented below is deprecated because the package architecture changed!). Here you see the code for creating such plots. Note that saving/displaying the results can take longt time. I prefer to render as mp4, but for this README a gif file was created.

Stephan_Avenue_Ped_daily <- get_EcoCounter_data(Stephan_Avenue_Ped, 
                                                from = "20141023", 
                                                to = "20181210", 
                                                step = 2) %>% 
  mutate(weekday = wday(date, label = TRUE),
         week = week(date),
         day = day(date),
         monthweek = ceiling(day(date) / 7),
         month = month(date, label = TRUE),
         year = year(date),
         hour = hour(date),
         minute = minute(date),
         time = strftime(date, format = "%H:%M")) %>% 
  na.omit()

g <- ggplot(Stephan_Avenue_Ped_daily, aes(monthweek, weekday, fill = comptage, frame = time)) + 
  geom_tile(colour = "white") + 
  facet_grid(year~month) + 
  scale_fill_gradient(low = "white", high = "red") +
  labs(x = "Week of Month",
       y = "Day",
       title = "Pedestrains at Stephan Avenue",
       fill = "Pedestrains")

library(gganimate)

## note: This process can take a long time!
gganimate(g, interval = 0.25, filename = "Pedestrains_Stephan_Avenue.gif")

Related projects