Wandering the streets of London after a long flight and basically no sleep in 36 hours, we were suffering from serious decision fatigue. Apps like Foursquare and Yelp are great to research possible places to go, but sometimes you just want to have a successful wander.
What does wandered.space do?
Wandered.space finds the nearest locations around you, and recommends a location to journey to. We consider ourselves “Experience-as-a-Service”. These locations are varied, including parks, libraries, restaurants, bars, cultural sites, and more. We’ve added a weight to a variety of location categories, allowing us to finetune what kinds of locations are suggested. Eventually, we will allow user input to dictate the growth of their own “journey” experience, through like/dislike prompts.
The unknown is often ephemeral; it slips through your hands, like the current of a tenuous stream. Wandered.space unlocks the potential of the unknown; it lets you take an adventure, wherever you are. Step-by-step, you’ll adventure your way through culture, history, and the natural world around you. Take a journey, wherever you may be.
How’d we get it done?
We built the UI using ReactJS and ArcGIS. The front-end consumes a series of endpoints that were created using the Azuqua Platform (a microservice application environment). Behind the scenes, we are looking up location data via the Factual API, and calculating distances between various points using the ArcGIS API. In tandem, this allows us to pinpoint the exact times between distances.
Hit up against the Factual API rate limit. Had to proxy our requests in RunKit (at 4AM) to get our service back up and running.
Additionally, we had to rollout from scratch a one-legged OAuth 1.0a solution for use on the Azuqua Platform.
- Awesome UI using ArcGIS.
- Insane OAuth1 implementation inside Azuqua.
- Finished on-time and delivered on what we were hoping to.
Learning The Hard Way
Don’t roll out your own solutions from scratch despite how fast you think you can do it. Don’t make too many requests to an API right before a demo at a hackathon.
What’s next for us?
We’d like to polish up the UX (we have big plans for that), finetune the algorithm, implement machine learning/predictive modeling on the the historical adventure data, allowing us to come up with the most interesting/well received adventures.