How an API can help improve marketing campaigns - An interview of Adrian Lopez, from Path travel

On May, 21st 2015, PredicSis and ChurnSpotter participated in the PAPIs Connect conference, an international event on machine learning and APIs, intended for decision makers, and held in Paris. The main focus of the event was to present to decision makers, data scientists, marketing managers and machine learning noobs, the latest technologies in predictive analytics, and the proeminent startups in the field.

As a way to promote our API and to allow people to discover PredicSis' products and to promote PAPIs Connect, we thought of a fun way to introduce attendees to the world of predictive analytics. We launched a contest.

The contestants were tasked with finding a secret sentence by using our API with the dataset we provided. Winners would be rewarded fancy prizes, such as an AppleWatch , vouchers, T-shirts…

The response was truly great and we got invaluable remarks and feedbacks on the API. Therefore we thought that it would be interesting to make a little interview with our winner, Adrián López Villa, CTO of PATH Travel, and learn a bit more about the business applications he envisions for APIs like PredicSis API.

So without further ado, here is the interview!

Adrián, can you tell us a bit more about what is PATH Travel about?

PATH Travel is a startup launched in 2014 that aims to build the future of mobile travel apps, by becoming a real travel assistant to businessmen and companies.

The idea is to aggregate, in a single location, the information received from different sources related to travels. Generally speaking, a traveler needs to constantly check flight times, and airport and weather updates. For a professional travel, you need to be particularly organized and thus be aware of traffic information, hotel reservations…etc. In addition, you have to be able to react rapidly and efficiently if plans change, or if you encounter any sort of problem that might impact your appointments.

But working closely with Leisure and Business Travel actors showed us that those who travel for business don’t always have time to keep in touch with all this information! So the idea for PATH Travel was born. Considering that the product is an app, it is easy for our customers to find all the relevant information about their travel in one place, on their phone and tablets.


So why did you come to PAPIs Connect exactly? What is the link between information aggregation and machine learning, in the context of PATH Travel?

We came to PAPIs Connect because we’d like to expand the services we offer to our clients. Presently, we aggregate all the information relevant for our customers’ travel. But we’d like to go beyond that and offer services based on our other clients suggestions, or our customer past choices. For example, we’d like to be able to use all the data we have on our clients choices to suggest to them hotels, means of transportation…etc.

But we’d also like to think ahead, in their stead. Let’s take an example. A client has had 3 appointments in Paris, from London, in the last 2 months. Each time, he took the same train, and the same taxi service to go to his appointment and everything was fine (thanks to our PATH Travel app, of course ;) ). But this time, thanks to the information that we have on the weather, traffic, or holidays in Paris, we know that using the same travel plans will take 2 more hours than usual for our client to reach his destination. Thus, using this knowledge on our client’s habits, and the additional information received from traffic…etc, we could be able to suggest him another departing time, or a motorcycle service instead of taxi, or even, a closer hotel available.

The idea, when coming to PAPIs Connect was that we wanted to be able to use all the data on our clients, the returns from our community, and additional information, to improve our clients’ travel experience, anticipate their needs and offer tailored alternatives.


Did you try other APIs, before testing PredicSis API through our PAPIs’ game?

We did. We use APIs daily to integrate new providers and content in our platform. Some time ago, we started working on predictive algorithms and trying some APIs. But it wasn’t what we expected, and it didn’t work the way we wanted it to.

So the idea behind our coming to PAPIs, was truly to find a new API that we could try out; or even, new ideas and technologies in machine learning that we could use to develop PATH Travel.


Seems like you enjoyed the game. Could you tell us how you found the answer?

Yes, I enjoyed it, even if I thought at first that I wouldn’t be able to use the ML Studio, considering that I’m a noob in this area. But, for the how I did it, quite easily frankly.

I got the contest table, shuffled it and took half the elements. I started using the Ruby SDK since I am more confident in Ruby, but I got some troubles so I moved to plain CURL requests. After a while I used the Python SDK, because editing CURL POST parameters was taking a lot of time.

Everything worked from there, and even if I was fighting a little bit with your API to understand what I could use, I managed to create a dataset, dictionary, set a target and create model, and finally to score it. And then, tada, the answer appeared: “PredicSis API is the answer”.

I got a cool AppleWatch while discovering technology that will help develop my company so, it was a good experience.


Well thanks for your answers and all the returns you’ve made on what we need to improve.