Flask petrol map

Initial setup

To setup the application to run on a new machine:

1. Create a virtual environment from the requirements.txt file
2. Copy .env.template to create an .env file and follow the instructions to create a secret key for the application along with a bing maps API key.
3. In the command line run flask run to start up the application.
4. Navigate to localhost:5000 and you should be able to see the live application.

Application purpose

The purpose of this application is for the user to enter their postcode and for the application to identify nearby petrol stations (within 2km). These petrol stations will be identified by their name and address in a table and will also be represented on a map to show which station may be closest to the users home.

Application implemenetation

Input form

Users are expected to input their home postocde (or current location postcode) and what fuel type they would like to search for. The Input for this is handled by a Flask-wtform which deals with UK postcode validation and the selection of fuel type.

Map representation

A interactive map is generated using the folium application. This is then passed to the user interface in the form of an IFrame object. The usage of a folium map object allows for the addition of interactive markers to show both the original home location and the petrol stations.

GoeCoding postcode

Geocoding the postcode is done using the BING API as integrated into the Geopy Geocding module. The reason for this is that it is much more likely to retrieve a point location for a postcode than the OSM API. It is also much cheaper than the Google Geocoding API. This point then location can then be passed to the OSM API to extract petrol stations nearby.

Getting petrol stations

Petrol stations are identified using the "amenity:fuel" tag from OSM and the "fuel":type key as well. The OSM API is used, through the OSMNX library as this is a freely available resource with no limits in terms of cost. For this the geometries_from_point function is used, as opposed to the geometries_from_address function because it is more reliable. As long as the postcode has been geocoded correctly then it should be able to find local petrol stations.

Mapping results

Results are mapped using the folium map object. This is done in three main stages:

1. Petrol stations are mapped onto the folium map object with markers
2. Home location is mapped onto the folium map object with a marker
3. A distance circle is created around the home location to show the distance boundary

Application workflow

The user will interact primarily with the input form where they input their postcode and fuel type. The application will then load the folium map object showing the result.

The index page looks like:

Where on first load the user sees the jumbotron, the input form and an empty map.

The user will then input their postcode. The validation of the postcode comes from the form which uses REGEX to determines whether the postcode follows the expected format of the UK postcode:

If the postcode is recognsied then the service will then attempt to geocode the postcode and find nearby petrol stations. If petrol stations are found then the result will show on the map along with a table showing the stations and their address:

ToDo

There project can be extened in several different ways. The way in which the project is currently build would allow for the adaptation towards finding other closest amenities in the future with a bit of refactoring. However, Keeping in line with the petrol station focus a few additions could inclide:

Adding petrol station distance

At the moment the application simply returns the petrol station name and address (if address is available). The user however is likely to want to know which of these stations is closer. The application could thus be extended to allow for the calculation of distance between the home loctaion and each petrol station, with an additional column added to the table to show distance.

Getting petrol station address

At the moment the address for the petrol station depends on the response from the OSM API. This can lead to missing information on the petrol station address even though it has geocordinates to say where it is. The application could thus be extended to make this search more robust.

Deploying application

At the moment the application needs to be deployed on individuals machines rather than being available to all users. It would be useful to deploy this application somewhere in the future. This could include deployment on an AWS lightsail instance of a Lambda service.

Adding Google Maps link

In addition to making the addresses more robuts, it would be beneficial for the user to be able to generate routes to the destination. This could be integrated through a google maps link on each address. This would allow the user to take advantage of google maps to find a route to the nearest or most appropriate destination.