search

MichaelsJP / Unrelevant

0 stars 0 forks source link

readme

Unrelevant Proof-of-Concept

Concept

Universal non-exclusive riser (to) elevate visualization accessibility (and) nescience (Unrelevant). Gives you functionality you never knew you could need. Enjoy! This tool is an extensive combination of popular open-source tools, namely openrouteservice and ohsome. The goal is to calculate accessibility statistics towards specified osm tags combined to coherent categories by using Isochrones from different routing providers with either time or distance as a range type. City boundaries and the osm tag data is crawled via the Ohsome API.

When using non-local based routing APIs be advised that depending on the size of the research area the amount of isochrones can easily reach five or more digits and exceed your quota without even finishing one run.

The current implementation is targeted to analyse accessibility to Green and Recreational POIs in cities in Germany but could easily adjusted to different target areas. The complete workflow can be seen in the following workflow:

unrelevant workflow

The complete research paper including explanations towards implementation can be found here: OPENSTREETMAP BASED ACCESSIBILITY ANALYSIS OF URBAN GREEN AND RECREATIONAL SPACES BY WALKING AND CYCLING FOR TEN CITIES IN GERMANY

For example results see chapter Exemplary Results

Dependencies

Installation

Install Poetry Unix/Linux

curl -sSL https://raw.githubusercontent.com/python-poetry/poetry/master/get-poetry.py | python -

Install Unrelevant (linux)

Clone and install locally:

# Create and activate virtual env, e.g.
python -m venv .venv
source .venv/bin/activate
poetry install --no-dev

See Usage or examples for more details.

Prepare the needed data sets

First download all the needed data sets that are too big to be served in this repository:

- The worldwide population data set: https://cidportal.jrc.ec.europa.eu/ftp/jrc-opendata/GHSL/GHS_POP_MT_GLOBE_R2019A/GHS_POP_E2015_GLOBE_R2019A_4326_9ss/V1-0/GHS_POP_E2015_GLOBE_R2019A_4326_9ss_V1_0.zip
- An osm.pbf for your desired spatial extent from: [http://download.geofabrik.de/europe.html](http://download.geofabrik.de/europe.html)

Then unzip the population data in a way that the .tif file is available.

Prepare the docker-compose setup

After the data download prepare the docker-compose.yml setup:

1. Replace the "./data/germany.osm.pbf" in  "- ./data/germany.osm.pbf:/ors-core/data/osm_file.pbf:ro`" with your chosen .osm.pbf.
2. If you're using a local ohsome instance uncomment the following:
#    environment:
#      DATA_FILE: "ohsome.oshdb.mv.db"
3. Replace the "ohsome.oshdb.mv.db" with the extract you downloaded.

Setup script

Run the setup script:

chomd +x setup.sh
./setup.sh -p ./your_population.tif

This will take care of the docker-compose setup. The bigger the population tif and osm.pbf for openrouteservice is, the longer the preparation will take. To see individual container output you can run one of the following comands:

docker logs unrelevant-ohsome-api -ft
docker logs unrelevant-ors-app -ft
docker logs unrelevant-postgres -ft
docker logs unrelevant-pgadmin4 -ft

Adjust the config.ini

Before you can run the tool, the config.ini needs to be checked. If nothing was changed in the docker-compose setup except the steps before, it should work out of the box. Else adjust the individual parameters in the config.ini file.

Tool execution

Once the setup is done, the tool is ready to be started. Run it with:

source .venv/bin/activate
python runner.py -c your-config.ini

Depending on the size of the input data the execution can take a lot of time. Be patient ;). The results will be worth it (hopefully).

Providers

Unrelevant supports three providers:

- valhalla (Open-Source)
- ors (Open-Source)
- here (Closed-Source/Proprietary)

The main goal is to achieve comparability among routing APIs from open-source projects. To have a reference from a routing engine, that is based on quality road data, here has been integrated. This gives the opportunity to back test results from free open-source projects based on OSM data, with results from data that was collected and prepared under paid market conditions.

Scenarios

Unrelevant supports the following scenarios:

- recreation

Usage

After installation, unrelevant can be called from the commandline as executable:

python runner.py --help
python runner.py --version
python runner.py -c config.ini

Config file parameters

The following parameters can be configured via the config.ini in the root folder.

[DEFAULT]

Scenario

Defines the scenario, at the moment only recreational.

Provider

Defines the desired routing provider.

- valhalla (via MapBox API)
- ors (via OpenRouteService API)
- here (via Here API)

Profile

Defines the desired profile that the isochrones should be generated with.

- bicycle
- pedestrian

Range_Type

Defines the range type. For valhalla only time is valid.

- time
- distance

Ranges

Defines the desired range in seconds or meters, depending on Range_Type.

Output_Folder

Defines the output folder where the results should be written to.

Verbosity

Defines the verbosity for the command line. Default is info.

Cities

Defines the list of cities with bounding boxes for faster ohsome querying. e.g.:

{
     "Berlin": "13.08835,52.33826,13.76116,52.67551",
     "Hamburg": "8.1053,53.3951,10.3253,54.0277",
     "München": "11.2809,48.0167,11.8247,48.3043",
     "Köln": "6.7725,50.8304,7.162,51.085",
     "Frankfurt am Main": "8.471825,50.015543,8.800472,50.227141",
     "Stuttgart": "9.038601,48.692019,9.31582,48.866399",
     "Düsseldorf": "6.688814,51.124375,6.939885,51.352532",
     "Leipzig": "12.236652,51.23817,12.542441,51.448114",
     "Dortmund": "7.302442,51.415525,7.638157,51.600041",
     "Essen": "6.894344,51.347571,7.13765,51.534202"
}

The ohsome query is boundary=administrative and name={city_name}. {city_name} stands for the iterative expansion with the names from the list. Make sure that the city names are actual names in OSM.

Threads

Defines the number of threads used for certain processes. Threads

Tags

Defines the list of categorized tags:

{
  "greenAreas": {
       "leisure":"park",
       "landuse":"grass",
       "landuse":"recreation_ground",
       "landuse":"meadow"
   },
   "water":{
       "leisure": "swimming_area",
       "natural": "beach",
       "leisure": "beach_resort",
       "water": "lake",
       "water": "reservoir",
       "water": "pond"
       },
   "historic":{
       "historic": "monument",
       "historic": "ruins",
       "historic": "castle",
       "historic": "fort",
       "historic": "memorial"
   },
   "tourism":{
       "tourism": "viewpoint",
       "tourism": "attraction",
       "natural": "peak",
       "tourism": "picnic_site",
       "leisure": "firepit",
       "amenity": "bbq",
       "amenity": "biergarten",
       "tourism": "camp_site",
       "tourism": "camp_pitch",
       "tourism": "caravan_site"
   }
}

[openrouteservice]

URL

Define the service URL.

Api_Key

Define the API key for public API access. Only needed when the url ist targeted at the public openrouteservice URL.

[Here]

Api_Key

[Valhalla]

Api_Key

Defines the valhalla api key for the public api. Not needed when a local instance is used.

https://openrouteservice.org/dev/#/signup
https://account.mapbox.com/access-tokens/

URL

Define the service URL.

[ohsome]

URL

Define the service URL.

[postgres]

URL

Define the service URL.

Port

Define the postgresql port

User

Define postgresql database user.

Password

Define postgresql database password.

Database

Define postgresql database name.

Execution example:

To run the example run from the root folder with a prepared environment:

python runner.py -c config.ini

Adjust the config.ini according to your needs.

Exemplary Recreational Results

The recreation scenario queries possible "Green and Recreational POIs" from the OHSOME API and calculates isochrones from each location. The isochrones represent a theoretical spatial coverage for the accessibility of the surrounding population. The population is calculated for each category based on the isochrones. The irochrones can be calculated using distance or time. For the valhalla routing engine only time is possible since distance is only implemented in the OpenRouteService engine.

config.ini parameters

[DEFAULT]
Scenario = recreation
Provider = ors
Profile = bicycle
Ranges = [150, 300, 450]
Range_Type = time
Output_Folder = ./output
Verbosity = info
Cities = {
         "Berlin": "13.08835,52.33826,13.76116,52.67551",
         "Hamburg": "8.1053,53.3951,10.3253,54.0277",
         "München": "11.2809,48.0167,11.8247,48.3043",
         "Köln": "6.7725,50.8304,7.162,51.085",
         "Frankfurt am Main": "8.471825,50.015543,8.800472,50.227141",
         "Stuttgart": "9.038601,48.692019,9.31582,48.866399",
         "Düsseldorf": "6.688814,51.124375,6.939885,51.352532",
         "Leipzig": "12.236652,51.23817,12.542441,51.448114",
         "Dortmund": "7.302442,51.415525,7.638157,51.600041",
         "Essen": "6.894344,51.347571,7.13765,51.534202"
         }
Threads = 10
Tags = {
   "greenAreas":
       {
       "leisure":"park",
       "landuse":"grass",
       "landuse":"recreation_ground",
       "landuse":"meadow"
   },
   "water":{
       "leisure": "swimming_area",
       "natural": "beach",
       "leisure": "beach_resort",
       "water": "lake",
       "water": "reservoir",
       "water": "pond"
   },
   "historic":{
       "historic": "monument",
       "historic": "ruins",
       "historic": "castle",
       "historic": "fort",
       "historic": "memorial"
   },
   "tourism":{
       "tourism": "viewpoint",
       "tourism": "attraction",
       "natural": "peak",
       "tourism": "picnic_site",
       "leisure": "firepit",
       "amenity": "bbq",
       "amenity": "biergarten",
       "tourism": "camp_site",
       "tourism": "camp_pitch",
       "tourism": "caravan_site"
    }
}

#### Results
Description Map Results
Category Green Areas with 300 second Isochrones from openrouteservice with the profiles food-walking and bicycle. Extract from the city of Munich: munich 300 seconds isochrone walking bicycle
Category Historic with 300 second Isochrones from openrouteservice with the profiles food-walking and bicycle. Extract from the city of Stuttgart: stuttgart 300 seconds isochrone walking bicycle
Category Tourism with 300 second Isochrones from openrouteservice with the profiles food-walking and bicycle. Extract from the city of Munich: munich 300 seconds isochrone walking bicycle
Description Statistical Results
Count POIs absolute pois abosoute
Count POIs per category pois per category per city
Profile cycling 300 seconds results per category and city 300 seconds isochrone bicycle
Profile walking 300 seconds results per category and city 300 seconds isochrone walking
Profile cycling 450 seconds results per category and city 450 seconds isochrone bicycle
Profile walking 450 seconds results per category and city 450 seconds isochrone walking