1 of 9

Flights Price Prediction

Using Supervised Learning

Presented By:

Deepika Goel

Praneeth

Under the Guidance of

Yasin Shah (CEO)

Technocolabs

2 of 9

SCRAPPED REAL-TIME DATA

First, try to scrap the data using the “travel payouts API” but was not getting all the information from the data.
Later, used web-scraping using “Skyscanner Flights API” but it also not contains the complete information.
So, we used “Selenium automation” and “Beautifulsoup” to scrape the real-time data from “Makemytrip.com”.
We have scrapped the data for “Economy class”, “Premium Economy class”, “Business class”.
We have scrapped the data in 15 days as it totally depends on the processor’s speed.

EXPLORATORY DATA ANALYSIS

We have first completed the analysis of the data while removing the duplicate values, null values and incorrect data that was fetched during the scrapping process.
Next process, checked the descriptive statistics, information that our data contains.
Removed the negative values from our data that we get while trying to calculate the difference between optimum time values.
Plotted different graphs like Bar Charts, Histograms, Box Plots, Cat Plots, Pie-chart and many more for Data Visualization.
Also, find out the feature extraction from our Data that how closely values are correlated with our dependent datasets: “Optimum Time” and “Price” values.

MODELLING PART

We have used Light GBM Regressor model to predict “Optimum Time” and “Price” on the basis of the cabin classes such as “Economy” and “Premium Economy”.
We have used CatBoost Regressor model to predict “Optimum Time” and “Price” on the basis of the “Business” cabin class.
So, total 6 models we had used. 3 models to predict “Optimum Time” first and remaining 3 models to predict the “Price”.

OUTPUT

DEPLOYMENT

We have used the “Streamlit” framework to deploy our models.
We have also linked “Research Paper” used for study, “Tableau Dashboard”, “Github Repository”, “Linkedin” profile, “Project Architecture” links for reference.
We had made the drop-down to select the “Departure Date”, “Departure City”, “Arrival City”, “Time” at which a customer wants to depart and on the basis of the selection our model will suggests the “Airlines” that has the “Price” with respect to the “Optimum Time”.
Customers can switch the view of “Price” in the output on the basis of the filter that is provided on the “Left-Hand” side of the “Dashboard” i.e. they can switch from “Low-High” and “High-Low”.
Created Map that will update the marker as per the arrival city, tooltip will show the name of the “Departure City” to “Arrival City” and pop-up will show the name of the “Departure City”.

DASHBOARD VIEW ON LOCAL MACHINE

DASHBOARD VIEW AFTER DEPLOYMENT

DEMO

Short video provided for the Dashboard/App that we created: