Below python script queries the OHSOME interface for the length of sidewalks mapped by different methods.

[Update: Correct accounting for sidewalk=both]

#!/usr/bin/python3

######### Plot the length of sidewalks mapped in certain regions

import json
import requests
import pandas as pd
import matplotlib.pyplot as plt
import datetime

pd.set_option('max_colwidth', 100) # maximum columnwidth for pandas
TIME_MONTHLY = "2008-01-01/2020-12-31/P1M" # temporal filter

######### Boundaries # http://norbertrenner.de/osm/bbox.html
BBOXES = {
    "US":                "-128.4,24,-57.2,49.7",
    "Poland":            "14.55,49.47,24.04,54.85",
    "Germany":           "5.89,47.17,13.86,54.89",
    "Great Britain":     "-11.7,49.8,2.1,61",
    "World":             "-180,-60,180,80", # took 1:10 min:s
    "East Coast":        "-83.7,23.9,-66.4,47.4",
    "Upper East Coast":  "-78.49,36.97,-70.11,42.94",
    "Not North America": "-25,-50,180,80|-110.6,-55.7,-29.2,24.5",
    "South America":     "-110.6,-55.7,-29.2,24.5",
}
BBOX="South America" # selected boundary

######### Filters
filterSidewalk = 'type:way and (\
(sidewalk=right) or \
(sidewalk=left) \
)'
filterSidewalk2 = 'type:way and (\
(sidewalk=both) \
)'
filterFootway = 'type:way and (\
(footway=sidewalk) \
)'

######### Query and plot
res = requests.post("https://api.ohsome.org/v1/elements/length/ratio",
      data = { 'filter': filterSidewalk, 'filter2': filterFootway,
               'bboxes': BBOXES[BBOX], 'time': TIME_MONTHLY } )
#print(res.text)
print(res.elapsed)
body = res.json()
df = pd.DataFrame(body['ratioResult'])
df.drop(columns=['ratio'], inplace=True)
# length is coming in meters, plot km
df.value = df.value/1000; df.value2 = df.value2/1000
df.timestamp = pd.to_datetime(df.timestamp)
df.set_index('timestamp', inplace=True)
res2 = requests.post("https://api.ohsome.org/v1/elements/length",
       data = { 'filter': filterSidewalk2,
                'bboxes': BBOXES[BBOX], 'time': TIME_MONTHLY } )
#print(res2.text)
print(res2.elapsed)
body2 = res2.json()
df2 = pd.DataFrame(body2['result'])
# length is coming in meters, plot km*2
df2.value = df2.value/500
df2.timestamp = pd.to_datetime(df2.timestamp)
df2.set_index('timestamp', inplace=True)
# plot result
df3 = df.add(df2, fill_value=0)
df3.rename(columns={'value': "Sidewalks", 'value2': "Footways"}, inplace=True)
df3.plot(title="Sidewalks (km) in OpenStreetMap in " + BBOX, figsize=(9.6, 4.8))
plt.show()

The resulting plots show, that there are indeed strong regional preferences:

Plots of the US and Eastern Europe look much like Poland; Germany and Western Europe look much like Great Britain. Plots with total number of ways have been created but not uploaded. The curves closely mirror the ones for length, when not accounting for sidewalk=both; their y-scale count slightly crosses the 100k mark, so average “raw” way length in both countries is roughly equal. If there is splitting going on, it affects both countries the same; When correctly accounting for sidewalk=both, average way length is considerably higher in Great Britain.

[Continue reading „Sidewalks by popularity“]