Customer Query - What are the coolest places to visit during my stay at Dallas?
Customer Query - Which is the nearest airport from where I am located in Chicago?
Customer Query - I want to fly from Bangalore for under $400, help!
Word Cloud with frequency of cities searched
With the advent of bots, we are utilizing AI/ML to accomplish tasks at a scale that is not matched by a team of humans. However, there is still a substantial amount of people who say that their conversation with the bot is unsuccessful. SkyMaster uses NLP which paves the way for easy-to-use interface and gives the end user on the other side of the screen the experience of having a conversation, as opposed to going through a limited set of options and menus to get the answer to their question.
What it does
Need to track flights in more detail than what the airport screens are telling you? Need a smart assistant for suggesting nearby tourist attractions during your next travel? Need a recommendation regarding the available airports nearby? Need someone to make you feel at home? SkyMaster does all of that for you. Above all, SkyMaster uses Natural Language Processing to ease user experience.
How we built it
We have used the following platforms:
- ReactJs for front-end
- Backend uses Flask and the APIs given by the Amadeus
- Google Cloud
- Tableau for Data Visualization.
SkyMaster is not exactly a chat-application; but somewhat similar to a bot. A search query of the user is transferred to DialogFlow, which extracts only the required keywords by using Natural Language Processing and passes it to Amadeus API which outputs the result. The output is fetched from flask server and populated on the front-end which uses ReactJs.
Challenges we ran into
- Amadeus API has a time-limit of only 30 minutes, making it difficult to integrate with the chat bot which uses URL en-coding.
- Mapping the city locations with the airport codes in DialogFlow took away major part of our time.
Accomplishments that we're proud of
- Overcame the challenge by creating the server on our own which accesses the API
- Efficient use of Amadeus API
- Staying up for two days at a stretch to build something that we designed.
- Completing the prototype with a working demo before the deadline.
- Efficient team co-ordination
What we learned
- ReactJs and DialogFlow integration
- Positive Attitude
- Time management
What's next for SkyMaster
Scalability, customization and a more better user interface.