Finnair needs to attract customers from emerging markets and give added value to transit customers. Some customers get premium experiences, but other customers could use more personalized information and guidance during their stay in Helsinki or during Transit. We have decided to build a scenario that nudges them towards new destinations based on customer profiling.
What it does
The web application TransPlanner is aimed at the current and potential transit passenger via social media advertisement. The application includes additional information about their transit point, and we analyze the social media profile to do a similarity comparison and to offer personalized suggestions for travel locations based on friend preferences, active events and popular destinations among the target cluster.
How we built it
Out team used promising machine learning approach to cluster facebook users.
Regarding the UI part, we employed Angular5 for views and front-end logic, Node.js to serve these views, GoogleAPI to show the results on the map, FacebookGraphAPI to login, Sketch for building the UX, Python for data preprocessing.
Challenges we ran into
We had different views of the product and a variety of ideas which were hard to combine into a unified vision. Development was also hard to integrate together due to different services.
Accomplishments that we're proud of
We did combine our ideas in a somewhat new concept.
What we learned
We gained a lot of experience with JS techology stack, as well as lean methodology of project design. We also learned to analyse Finnair's problems and possibilities to solve them.
What's next for Finnair TransPlanner
We'd love to finish it into a polished product and start a discussion to continue collaboration with the company.