The theme of the Emirates CMU Innovation Lab is inspiring insights, personalized customer experiences and data analytics driven by machine learning and AI. They providing data from Emirates/JetBlue airline and other publicly available sources, and the objective for our participants is to use their ingenuity to create innovative and workable solutions that align with the theme
We believe we could use Machine Learning to process/analyze the users' trip and register information, to determine the latest city the users is without having s/he explicitly update.
What it does
It analyze the mock up contact/trip data and come up the best reliable city location of the user.
How we built it
We first use SalesForce Einstein to review/model the data, then we written the code in Python and ran the result using Intel-Nervana platform.
Challenges we ran into
We were originally attempt to tackle two issues (the right pricing), but the we pivot back to the location context.
Accomplishments that we're proud of
We were able to predict 39% of the user could have their home city updated.
What we learned
Modeling using Temporal analysis and Graph analysis through machine learning in this unsupervised learning that does not have a final answer.
What's next for DataOnTheFly - Emirates CMU ML and AI Travel Hackathon
We might attempt to review the rest of the data and perhaps add the Social Media to help re-inforce the location accuracy.