A Tidy Interface to the ‘Walk Score’ API

This package provides a tidy interface to the Walk Score API, a proprietary API that measures a location’s “walkability” using a number between 0 and 100.

The Walk Score API has a free tier which allows 5,000 API calls per day, and paid tiers with higher limits.

This function makes it easy to spread your API calls out over a few days. When you call the function for the first time, if necessary it creates a new column of walks cores and assigns each row NA. Then, each row’s walk score is populated as the function gets a good API response. The function breaks automatically upon detecting a rate limit, returning all results collected so far. When your rate limit resets and you call the function again, it picks up from the first NA walk score it finds and continues on. So make sure to save your results after each batch, but you don’t need to keep track of fine-grained batch issues or worry about losing a whole batch if a response errors out–the function handles that for you.

You’ll need a valid Walk Score API key to use this package.

Please Note neither this package nor its author are affiliated with Walk Score in any way, nor are any warranties made about this package or any data available through the Walk Score API. “Walk Score” is copyrighted and a registered trademark of its owner, again, with whom we are not affiliated.

API documentation is available here: https://www.walkscore.com/professional/api.php

Installation

You can install the development version of walkscore like so:

devtools::install_github("https://github.com/chris31415926535/walkscore")

Example

library(dplyr)
library(walkscore)

your_apikey <- "put a real API key here"

test_data <- dplyr::tibble(lat = 45.420193, lon = -75.697796) |>
  walkscore::walkscore(apikey = your_apikey)

A Note about NA API Responses

If you receive many NA responses, try waiting a few minutes and repeating your query. When you query the Walkscore API for specific lat/lon value, it seems that it returns cached values if they exist, or otherwise calculates new values. However, in at least some cases, the API seems to return an invalid/nonexistent result to your query while it is calculating your requested value on the server.

So if you are querying regions that are new to Walkscore you will receive NA results the first time, but if you query again later Walkscore may have calculated those results and added them to its cache.

Unfortunately this is a server-side issue and this package can’t fix it: we can only parse the responses the server gives us, and the polite_pause parameter is for adding delays between API calls whatever their result may be.

Another reminder that I am not affiliated with Walkscore!

Related Projects

See also the package walkscoreAPI available on CRAN. Compared to walkscoreAPI, the current package walkscore has advantages for some use cases:

1. walkscore uses data frames for inputs and outputs and adheres to “tidy” design princniples; walkscoreAPI works on single values and provides output as a list.
2. walkscore handles batching automatically for data frame inputs; walkscoreAPI does not.
3. walkscore automatically handles API failures if you hit your rate limit by returning the results so far, and will pick up where it left off if you re-feed the output into it again once your rate limit resets; walkscoreAPI has no such functionality.

However, walkscoreAPI may be simpler if you only need to find a few values, or if you have a professional/enterprise API key with higher usage limits and you want to run a large volume of API calls in parallel.

walkscoreAPI also has a number of helper functions, whereas walkscore is focused entirely on accessing the Walk Score API.