Postal Code

It provides an api endpoint which takes a GPS latitude and longitude and spits out the names of museums around that location grouped by their postcode as JSON.

Assumptions:

Postalcode data from WhosOnFirst.org covers the countries fully for postal codes
Postalcode data hasn't been changed since they obtained the administrative & postcode data
Coverage is good enough for the domain area

Example result

As an example when doing a request to /museums?lat=52.494857&lng=13.437641 would generate a response similar to:

{
  "10999": ["Werkbundarchiv – museum of things"],
  "12043": ["Museum im Böhmischen Dorf"],
  "10179": [
    "Märkisches Museum",
    "Museum Kindheit und Jugend",
    "Historischer Hafen Berlin"
  ],
  "12435": ["Archenhold Observatory"]
}

Local setup

We need postgresql and redis to store data.

Preparing data

psql
# create user and add username and password to .env
CREATE ROLE postalcode WITH CREATEDB LOGIN SUPERUSER PASSWORD 'postalcode';

# ruby 2.6.3
rvm use 2.6.3
gem instsall bundler

cp .env.example .env
bundle install

rake db:setup

# Japen db by default
# We will import WhosOnFirst administrative and postcode data
rake wof:download_db
rake wof:convert_to_csv
rake wof:import_csv_to_db
rake wof:import_geojson_to_tile38

or

rake wof:setup # to run 4 steps above

# For adding extra migrations
rake db:new_migration name=CreateSomething

# In a new terminal, launch tile38
# https://tile38.com/topics/getting-started/
# tile38-server will use redis://localhost:9851/0 by default
# It will consume about 560MB ram.
tile38-server

# In a new terminal, launch ruby server
# And visit http://localhost:9292
bundle exec rackup

Usage examples

Museums

http://localhost:9292/museums?lat=52.494857&lng=13.437641

{
  "10963": [
    { "text": "Museumspark", "postcode": "10963" },
    { "text": "Museumsshop", "postcode": "10963" }
  ],
  "10785": [{ "text": "Museumsshop", "postcode": "10785" }],
  "12627": [{ "text": "Museumswohnung Hellersdorf", "postcode": "12627" }],
  "15562": [{ "text": "Museumspark Rüdersdorf", "postcode": "15562" }]
}

Starbucks Japan search

keyword is starbucks
coordinate is lat: 35.7020691, lng: 139.7753269

http://localhost:9292/starbucks?lng=139.7753269&lat=35.7020691

result is

{
  "101-0028": [{ "text": "Starbucks", "postcode": "101-0028" }],
  "110-0007": [{ "text": "Starbucks", "postcode": "110-0007" }],
  "101-0045": [{ "text": "Starbucks", "postcode": "101-0045" }],
  "110-0005": [
    { "text": "Starbucks", "postcode": "110-0005" },
    { "text": "Starbucks", "postcode": "110-0005" }
  ]
}

Form [35.7020691,139.7753269] to postal code location 101-0062. google shows that the distance is 1.6 km

Mapbox doesn't return postcode for that coordinate.