victor-savu / lunch-walk Goto Github PK

Provide a link with directions to the selected food truck. It may be an external link (e.g. to Google Maps)

Use the browser location as the starting point so that the user doesn't have to type in coordinates.

While in #4 we get accurate walking distances, as our customer base gets bigger/hungrier, we will make excessive calls to the Bing Maps API and run up a mighty bill.
Is there a way to buy the static maps data and implement some rudimentary query operations that would reduce our dependency on external APIs?

The original data from the city of San Francisco contains more information than we need.

Implement an adapter service that narrows the data down to the necessary columns.

Use a maps provider to get more accurate walking distances for ranking the food trucks.

Bing Maps seem to have a nice API for that, which might come in handy.

For the minimum functionality, present a web page with two input fields: latitude and longitude and a search button.

Display the 5 nearest food carts to the input location.

Approximation: use Euclidean distance and assume that latitude and longitude are Cartesian coordinates.

Note for later: This approximation, while good for a Proof of Concept, will break down pretty quickly. The Earth's curvature is negligible for our case, but elevation can change abruptly in San Francisco even over short distances, which affects walk times. Furthermore, the street layout will likely prohibit the user from walking in a straight line to any destination, so the Euclidean distance is an optimistic estimate.