FareBoom Price Predictor
Data Analysis , Data Mining, Data Analytics
FareBoom Price Predictor is designed for the price sensitive consumer to target the best possible time to buy an airline ticket over several months. While competitive solutions might predict fare movement up to 7 days in advance, Fareboom goes beyond this limitation that is impractical when considering a purchase several months in advance of trip departure.
Approach: AltexSoft’s Data Science team was given a challenging task, to create an algorithm that predicts whether airline prices will go up or down over the next seven days. This is a Time Series forecasting problem.
After the problem research, data analysis and a lot of experiments, the team came up with an algorithm that uses ARIMA for forecasting. Along with several additional data filtering, munging technics, and ensemble voting, it gave a correct prediction up to 85% cases based on validation testing.
Benefits: Creation of better retention programs to increase customer loyalty via additional features to predict the prices
Tools and algorithms: R (modeling), ARIMA, Ensembles, C# (implementation)
Cloud-based web services
, Data Analysis
, Data Analytics
, Data Mining
, Machine Learning
, Predictive Analytics
, Price Predictor