Skip to content

Latest commit

 

History

History
543 lines (395 loc) · 20.8 KB

CONTRIBUTING.md

File metadata and controls

543 lines (395 loc) · 20.8 KB
Raw

Contributing

tl;dr

🙋   How to help:
  • npm run build will reprocess the files and output warnings
  • Resolve warnings - show me
  • Remove generic names - show me
  • Add brand:wikidata and brand:wikipedia tags - show me
  • Add missing brands - show me
  • Edit Wikidata - show me

Tip: You can browse the index at http://osmlab.github.io/name-suggestion-index/brands/index.html to see which brands are missing Wikidata links, or have incomplete Wikipedia pages.

🚫   Don't edit the files in dist/ - they are generated:
  • dist/names_all.json - all the frequent names and tags collected from OpenStreetMap
  • dist/names_discard.json - subset of names_all we are discarding
  • dist/names_keep.json - subset of names_all we are keeping
  • dist/wikidata.json - cached brand data retrieved from Wikidata
✅   Do edit the files in config/ and brands/:
  • config/*
    • config/filters.json- Regular expressions used to filter names_all into names_keep / names_discard
  • brands/* - Config files for each kind of branded business, organized by OpenStreetMap tag
    • brands/amenity/*.json
    • brands/leisure/*.json
    • brands/shop/*.json
    • brands/tourism/*.json

 

Background

🗺️   About OpenStreetMap

OpenStreetMap is a free, editable map of the whole world that is being built by volunteers. Features on the map are defined using tags. Each tag is a key=value pair of text strings.

For example, a McDonald's restaurant might have these tags:

  "amenity": "fast_food"
  "cuisine": "burger"
  "name": "McDonald's"
  ... and more tags to record its address, opening hours, and so on.

 

💡   About the name-suggestion-index

The goal of this project is to define the most correct tags to assign to each common brand name. This helps people contribute to OpenStreetMap, because they can pick "McDonald's" from a list and not need to worry about the tags being added.

 

🗃️   About the brand files

The brands/* folder contains many files, which together define the most correct OpenStreetMap names and tags.

These files are created by a several step process:

  • Process the OpenStreetMap "planet" data to extract common names -> dist/names_all.json
  • Filter all the names into -> dist/names_keep.json and dist/names_discard.json
  • Merge the names we are keeping into -> brands/**/*.json files for us to decide what to do with them

The files are organized by OpenStreetMap tag:

  • brands/* - Config files for each kind of branded business, organized by OpenStreetMap tag
    • brands/amenity/*.json
    • brands/leisure/*.json
    • brands/shop/*.json
    • brands/tourism/*.json

Each brand entry looks like this (comments added for clarity):

In brands/amenity/fast_food.json:

  "amenity/fast_food|McDonald's": {         // Identifier like "key/value|name"
    "count": 19040,                         // `count` - generated # of these that we found in OpenStreetMap
    "match": [
      "amenity/fast_food|Mc Donald's",      // Optional `match` array
      "amenity/fast_food|McDonalds",        //   Contains less-desirable variations to ignore.
      "amenity/restaurant|Mc Donald's",     //   (we want to keep only "amenity/fast_food|McDonald's")
      "amenity/restaurant|Mc Donalds",      //
      "amenity/restaurant|McDonald's",      //
      "amenity/restaurant|McDonalds",       //
    ],
    "tags": {                               // "tags" - OpenStreetMap tags that every McDonald's should have
      "amenity": "fast_food",               //   The OpenStreetMap tag for a "fast food" restaurant
      "brand": "McDonald's",                //   `brand` - Brand name in the local language (English)
      "brand:wikidata": "Q38076",           //   `brand:wikidata` - Universal Wikidata identifier
      "brand:wikipedia": "en:McDonald's",   //   `brand:wikipedia` - Reference to English Wikipedia
      "cuisine": "burger",                  //   `cuisine` - What kind of fast food is served here
      "name": "McDonald's"                  //   `name` - Display name, also in the local language (English)
    }
  },

There may also be entries for McDonald's in other languages!

  "amenity/fast_food|マクドナルド": {         // Identifier like "key/value|name"
    "count": 1445,
    "countryCodes": ["jp"],                 // Optional `countryCodes` - array of countries where this entry is used
    "tags": {
      "amenity": "fast_food",
      "brand": "マクドナルド",                // `brand` - Brand name in the local language (Japanese)
      "brand:en": "McDonald's",             // `brand:en` - For non-English brands, tag the English version too
      "brand:ja": "マクドナルド",             // `brand:ja` - Add at least one `brand:xx` tag that matches `brand`
      "brand:wikidata": "Q38076",           // `brand:wikidata` - Same Universal wikidata identifier
      "brand:wikipedia": "ja:マクドナルド",   // `brand:wikipedia` - Reference to Japanese Wikipedia
      "cuisine": "burger",
      "name": "マクドナルド",                 // `name` - Display name, also in the local language (Japanese)
      "name:en": "McDonald's"               // `name:en` - For non-English names, tag the English version too
      "name:ja": "マクドナルド",              // `name:ja` - Add at least one `name:xx` tag that matches `name`
    }
  },

 

Optional properties to suppress warnings

These properties always start with "no".

nomatch

Sometimes there are multiple different entries that use the same name.

For example, "Sonic" can be either a fast food restaurant or a fuel station.

We want to allow both kinds of "Sonic" to exist in the index, and we don't want to be warned that they are potentially duplicate, so we can add a nomatch property to each entry to suppress the "duplicate name" warning.

  "amenity/fuel|Sonic": {
    "count": 127,
    "nomatch": ["amenity/fast_food|Sonic"],
    ...
  }
  "amenity/fast_food|Sonic": {
    "count": 1110,
    "nomatch": ["amenity/fuel|Sonic"],
    ...
  }
nocount

Sometimes a new, uncommon OpenStreetMap tag-name combination should replace an existing tag-name combination.

For example, many Costcos are tagged as shop=department_store or shop=supermarket, but the new preferred tag is shop=wholesale.

We don't want to be warned that the new correct tag is uncommon, so we add a nocount property to the entry to suppress the "uncommon name" warning.

  "shop/wholesale|Costco": {
    "nocount": true,
    "match": [
      "shop/department_store|Costco",
      "shop/department_store|Costco Wholesale",
      "shop/supermarket|Costco",
      "shop/supermarket|Costco Wholesale"
    ],
    ...
  },

 

Identical names, multiple brands

Sometimes multiple brands with the same name will operate in geographically distinct locations. You can modify the key to include a tilde ~ after the name to tell the difference between two otherwise identical brands. The text after the tilde can contain anything.

When using a tilde ~ name:

  • You should add "nocount": true, as the script will not be able to determine the true count.
  • You should add "nomatch": properties to each name so they do not generate duplicate name warnings.
  "shop/supermarket|Price Chopper~(Kansas City)": {
    "countryCodes": ["us"],
    "nocount": true,
    "nomatch": [
      "shop/supermarket|Price Chopper~(New York)"
    ],
    "tags": {
      "brand": "Price Chopper",
      "brand:wikidata": "Q7242572",
      "brand:wikipedia": "en:Price Chopper (supermarket)",
      "name": "Price Chopper",
      "shop": "supermarket"
    }
  },
  "shop/supermarket|Price Chopper~(New York)": {
    "countryCodes": ["us"],
    "nocount": true,
    "nomatch": [
      "shop/supermarket|Price Chopper~(Kansas City)"
    ],
    "tags": {
      "brand": "Price Chopper",
      "brand:wikidata": "Q7242574",
      "brand:wikipedia": "en:Price Chopper Supermarkets",
      "name": "Price Chopper",
      "shop": "supermarket"
    }
  },

 

What you can do

🏗️   Building the project

To rebuild the index, run:

  • npm run build

This will output a lot of warnings, which you can help fix!

 

🤔   Resolve warnings

Warnings mean that you need to edit files under brands/*. The warning output gives a clue about how to fix or suppress the warning. If you aren't sure, just ask on GitHub!

 

Duplicate names

Warning - Potential duplicate brand names:
To resolve these, remove the worse entry and add "match" property on the better entry.
To suppress this warning for entries that really are different, add a "nomatch" property on both entries.
  "tourism/motel|Motel 6" -> duplicates? -> "tourism/hotel|Motel 6"

What it means: "Motel 6" exists in the index twice - as both a tourism=hotel (wrong) and a tourism=motel (correct). In this situation we want to:

  • Delete the entry for "tourism/hotel|Motel 6" and
  • Add "match": ["tourism/hotel|Motel 6"] to the "tourism/motel|Motel 6" entry

Local knowledge, existing tagging (indicated by the count property), information at the relevant Wikipedia page or the company's website, and OpenStreetMap Wiki tag documentation all help in deciding which entry should be kept.

If the situation is unclear, one may contact the local community and ask for help.

Note that in some cases both entries should be kept - for example, a given brand may really operate both supermarkets and convenience stores under the same name. In that case it is necessary to use the nomatch property.  

Uncommon names

Warning - Uncommon brand not found in `names_keep.json`:
These might be okay. It just means that the entry is not commonly found in OpenStreetMap.
To suppress this warning, add a "nocount" property to the entry.
  "shop/wholesale|Costco"

What it means: This warning can occur if the index contains an entry that is not common in OpenStreetMap. Either replace it with a more common tag, or add "nocount": true to suppress the warning.

(In this situation, shop=wholesale is a new preferred tag, but most existing Costcos were still tagged as shop=department_store. Suppressing the warning is the correct thing to do).

 

🔪   Remove generic names

Some of the common names in the index might not actually be brand names. We want to remove these generic words from the index, so they are not suggested to mappers.

For example, "Универмаг" is just a Russian word for "Department store":

  "shop/department_store|Универмаг": {
    "count": 210,
    "tags": {
      "brand": "Универмаг",
      "name": "Универмаг",
      "shop": "department_store"
    }
  },

To remove this generic name:

  1. Delete the item from the appropriate file, in this case brands/shop/department_store.json
  2. Edit config/filters.json. Add a regular expression matching the generic name in either the discardKeys or discardNames list.
  3. Run npm run build - if the filter is working, the name will not be put back into brands/shop/department_store.json
  4. git diff - to make sure that the entries you wanted to discard are gone (and no others are affected)
  5. If all looks ok, submit a pull request with your changes.

 

🕵️‍♀️   Add wiki tags

Adding brand:wikipedia and brand:wikidata tags is a very useful task that anybody can help with.

Example #1 - Worldwide / English brands...

  1. Find an entry in a brand file that is missing these tags:

In brands/amenity/fast_food.json:

  "amenity/fast_food|Chipotle": {
    "count": 708,
    "match": [
      "amenity/fast_food|Chipotle Mexican Grill",
      "amenity/restaurant|Chipotle",
      "amenity/restaurant|Chipotle Mexican Grill"
    ],
    "tags": {
      "amenity": "fast_food",
      "brand": "Chipotle",
      "cuisine": "mexican",
      "name": "Chipotle"
    }
  },
  1. Google for that brand - if you are lucky, you might find the Wikipedia page right away.

Google for Chipotle

  1. From the Wikipedia page URL, you can identify the brand:wikipedia value.

OpenStreetMap expects this tag to be formatted like "en:Chipotle Mexican Grill".

  • Copy the page name from the URL.
  • Add the language prefix - "en:" for the English Wikipedia.
  • Replace the underscores '_' with spaces.

On the brand's Wikipedia page, you can also find its "Wikidata item" link. This appears under the "tools" menu in the sidebar.

👉 protip: @maxerickson has created a user script to make copying these values even easier - see #1881

Chipotle Wikipedia

  1. On the brand's Wikidata page, you can identify the brand:wikidata value. It is a code starting with 'Q' and several numbers.

Chipotle Wikidata

  1. Update the brand file, in this case brands/amenity/fast_food.json:

We can add the "brand:wikipedia" and "brand:wikidata" tags.

  "amenity/fast_food|Chipotle": {
    "count": 708,
    "match": [
      "amenity/fast_food|Chipotle Mexican Grill",
      "amenity/restaurant|Chipotle",
      "amenity/restaurant|Chipotle Mexican Grill"
    ],
    "tags": {
      "amenity": "fast_food",
      "brand:wikidata": "Q465751",                            // added
      "brand:wikipedia": "en:Chipotle Mexican Grill",         // added
      "brand": "Chipotle",
      "cuisine": "mexican",
      "name": "Chipotle"
    }
  },

(comments added for clarity)

  1. Rebuild and submit a pull request.
  • Run npm run build
  • If it does not fail with an error, you can submit a pull request with your changes.

 

Example #2 - Regional / non-English brands...

This example uses a brand "かっぱ寿司". I don't know what that is, so I will do some research.

  1. Find an entry in a brand file that is missing these tags:

In brands/amenity/fast_food.json:

  "amenity/fast_food|かっぱ寿司": {
    "count": 91,
    "tags": {
      "amenity": "fast_food",
      "brand": "かっぱ寿司",
      "name": "かっぱ寿司"
    }
  },
  1. Google for that brand - if you are lucky, you might find the Wikipedia page right away.

Tip: You might want to narrow you search by Googling with a site: filter: "かっぱ寿司 site:ja.wikipedia.org"

From these results, we can know that the brand is "Kappazushi", owned by a Japanese company called "Kappa Create". We can also find the Wikipedia page.

Google for かっぱ寿司

  1. As with English brands, you can identify the brand:wikipedia value from the URL. Because this is a Japanese brand, we will link to the Japanese Wikipedia page.

OpenStreetMap expects this tag to be formatted like "ja:かっぱ寿司".

  • Copy the page name from the URL.
  • Add the language prefix "ja:".
  • Replace the underscores '_' with spaces.

Although I can not read Japanese, I can identify the "Wikidata item" link because it always appears in the sidebar and mouseover will show the Wikidata 'Q' code in the URL.

Kappa Sushi Wikipedia

  1. On the brand's Wikidata page, you can identify the brand:wikidata value. It is a code starting with 'Q' and several numbers.

Note: The Wikidata page looks a bit sparse - you can edit this too if you want to help!

Kappa Sushi Wikidata

  1. Update the brand file, in this case brands/amenity/fast_food.json:

We can add:

  • "brand:en" and "name:en" tags to contain the English name "Kappazushi"
  • "name:ja" and "brand:ja" tags to contain the local name "かっぱ寿司"
  • "brand:wikipedia" and "brand:wikidata" tags
  • "cuisine": "sushi" OpenStreetMap tag
  • "countryCodes" property, to indicate that this brand should only be used in Japan.
  "amenity/fast_food|かっぱ寿司": {
    "count": 91,
    "countryCodes": ["jp"],                // added
    "tags": {
      "amenity": "fast_food",
      "brand": "かっぱ寿司",
      "brand:en": "Kappazushi",            // added
      "brand:ja": "かっぱ寿司",            // added
      "brand:wikipedia": "ja:かっぱ寿司",     // added
      "brand:wikidata": "Q11263916",        // added
      "cuisine": "sushi",                   // added
      "name": "かっぱ寿司",
      "name:en": "Kappazushi",              // added
      "name:ja": "かっぱ寿司"               // added
    }
  },

(comments added for clarity)

  1. Rebuild and submit a pull request.
  • Run npm run build
  • If it does not fail with an error, you can submit a pull request with your changes.

 

🏪   Add missing brands

If it exists, we want to know about it!

Some brands aren't mapped enough (50+ times) to automatically be added to the index so this is a valuable way to get ahead of incorrect tagging.

  1. Before adding a new brand, the minimum information you should know is the correct tagging required for instances of the brand (name, brand and what it is - e.g. shop=food). Ideally you also have brand:wikidata and brand:wikipedia tags for the brand and any other appropriate tags - e.g. cuisine.

  2. Add your new entry anywhere into the appropriate file under brands/* (the files will be sorted alphabetically later) and using the "tags" key add all appropriate OSM tags. Refer to here if you're not familiar with the syntax.

  3. Add the "nocount": true key to your new entry as this will suppress the warnings that it doesn't appear often in OSM.

  4. If the brand only has locations in a known set of countries add the "countryCodes": [] key to your new entry. This takes an array of ISO 3166-1 alpha-2 country codes in lowercase (e.g. ["de", "at", "nl"]).

  5. If instances of this brand are commonly mistagged add the "match": [] key to list these. Again, refer to here for syntax.

  6. Run npm run build and resolve any duplicate name warnings.

 

Using Overpass Turbo

Sometimes you might want to know the locations where a brand name exists in OpenStreetMap. Overpass Turbo can show them on a map:

  1. Go to https://overpass-turbo.eu/

  2. Enter your query like this, replacing the name and other OpenStreetMap tags. Because we don't specify a bounding box, this will perform a global query.

nwr["name"="かっぱ寿司"]["amenity"="fast_food"];
out center;

Tip: The browsable index at http://osmlab.github.io/name-suggestion-index/brands/index.html can open Overpass Turbo with the query already set up for you.

  1. Click run to view the results.

As expected, the "かっぱ寿司" (Kappazushi) locations are all concentrated in Japan.

Overpass search for かっぱ寿司

 

📝   Edit Wikidata

Editing brand pages on Wikidata is something that anybody can do. It helps not just our project, but anybody who uses this data for other purposes too! You can read more about contributing to Wikidata here.

  • Add Wikidata entries for brands that don't yet have them.
  • Improve the labels and descriptions on the Wikidata entries.
  • Translate the labels and descriptions to more languages.
  • Add social media accounts under the "Identifiers" section. If a brand has a Facebook, Instagram, or Twitter account, we can fetch its logo automatically.

Tip: The browsable index at http://osmlab.github.io/name-suggestion-index/brands/index.html can show you where the Wikidata information is missing or incomplete.

Adding information on Wikidata