Advancement and Research in Instrumentation Engineering

Volume 5 Issue 1
e-ISSN: 2582-4341

Flight Ticket Price Predictor Using Python

Vivekanand P. Thakare, Ankita Sanjay Murraya, Roshani Bandu Gawade, Mrunali
Mukundrao Sawarkar, Trupti Khemraj Shende, Ujjwala Kamlesh Badole
Department of Computer science & Engineering, GWCET, Nagpur, Maharashtra, India
Corresponding Author
E-Mail Id: [email protected]

ABSTRACT
Boxer timing for airline ticket purchasing from the consumer’s perspective is challenging
principally because buyers have insufficient information for reasoning about future price
movements. In this project, we mainly directed to uncover underlying drifts of flight prices in
India using historical data and also to advise the best time to buy a flight ticket. Remarkable,
the trends of the price are highly sensitive to the route, month of departure, day of departure
time of the day is a holiday and airline carrier. Highly competitive routes like most business
routes hand a non-decreasing trend where prices increased as day to departure decreased

Keywords: Feature selection, airfare price, prediction model, random forest, pricing models

INTRODUCTION flights even for nearby seats within the

For this project, we have gathered data same cabin [6,9], customers are seeking
from 18 routes all over India while the to get the lowest price while airlines are
data of 4 routes were broadly used for the trying to keep their overall revenue as high
analysis due to the sheer volume of data as possible and maximize their profit [3].
collected over 4 months bring about 5.28
lakh data points each across the Mumbai- Airlines use various kinds of
Delhi and Delhi-Mumbai route and 1.05 computational techniques to increase their
lakh data points crosswise the Delhi- revenue such as demand prediction and
Guwahati and Guwahati-Delhi route [1-4]. price discrimination. On the customer side
[10], two kinds of models are proposed
The project implements is the validations by different researchers to save money for
or contradictions towards myths regarding customers models that predict the optimal
the airline industry, a comparison study time to buy a ticket and models that
among various models in predicting the predict the minimum ticket price [4,5].
optimal time to buy the flight ticket and
the amount that can be saved if done so, A Time of purchase patterns[11] (making
customized model which included [8] a sure last-minute purchases are expensive)
combination of ensemble and statistical 2. Keeping the flight as full as they want it
models have been implemented with the (raising prices on a flight that is filling up
best accuracy of above 90% for a few to reduce sales and hold back inventory for
routes, mostly from Tier to metro cities those expensive last-minute expensive
[7]. These models have led to substantial purchases)
savings and formed average positive
savings on every transaction. If we could inform the travelers about the
boxer time to buy their flight tickets based
BACKGROUND on the historic data and show them various
Nowadays airline ticket prices can vary trends in the airline industry [11,12], we
dynamically and significantly for some could help them save money on their

HBRP Publication Page 1-5 2022. All Rights Reserved Page 1

 Advancement and Research in Instrumentation Engineering
Volume 5 Issue 1
e-ISSN: 2582-4341

travels. This would be a practical  Always available for making the

implementation of data analysis, statistics reservation.
[13], and machine learning techniques to  Cancelation and changes of travel.
solve a daily problem faced by travelers.
PROPOSED MODEL
LITERATURE SURVEY  After we need data, we need to clean
 K. Tziridis, T. Kalampokas It is hard & prepare the Data according to the
for the client to buy an air ticket at the model’s requirement.
most reduced cost. For this few  We used various statistical techniques
procedures are explored to determine & logic and implemented those using
the time and date to grab air tickets built-in packages.
with a minimum fare rate [12,]. The  In this paper [4], random forest and
majority of these systems are utilizing decision tree algorithms are to find
the modern computerized system solutions for flight price problems in
known as Machine Learning. To machine learning tasks, and a hybrid
determine the ideal purchase for a method is formulated from Chi-
flight ticket. square, ANOVA, and Correlation tests
 A. Sharma, since the deregulation of are performed, departure and arrival
the airline industry, airfare pricing features play an important role in
strategy has developed into a complex predicting the price. Running the
structure of sophisticated rules and random forest and decision tree
mathematical models that drive the models, grouping the maximum price
pricing strategies of airlines. of airlines. Next, perform the feature
 Hoseinzade E, Haratizadeh S. (2019) engineering and calculate the accuracy
Cnnpred: CNN-based stock market  In this article, we will be analyzing the
prediction using a diverse set of flight fare prediction using a Machine
variables. Expert Syst Appl Learning dataset using essential
129(SEP.):273–285Traditional exploratory data analysis techniques
variables such as distance, although then will draw some predictions about
still playing a significant role, are no the price of the flight based on arrival
longer the sole factor that dictates the time, what is the departure time, what
pricing strategy. Elements related to is the duration of the flight, source,
economic, marketing, and societal destination and more.
trends have played increasing roles in  Flight ticket prices can be something
dictating airfare prices. hard to guess, today we might see a
 T. Wang et al., "A Framework for price, check out the price of the same
Airfare Price Prediction: A Machine flight tomorrow, and it will be a
Learning Approach. A Linear Quantile different story.
Mixed Regression Model for  To solve this problem, we have been
Prediction of Airline Ticket Price [4]. provided with prices of flight tickets
for various airlines between January
ADVANTAGE and March of 2021 and between
 Travelers get the fare prediction handy various cities, using which we aim to
using which it’s easy to decide on the build a model which predicts the prices
airline. of the flights using various input
 Save time in searching / deciding for features.
the airline.

HBRP Publication Page 1-5 2022. All Rights Reserved Page 2

 Advancement and Research in Instrumentation Engineering
Volume 5 Issue 1
e-ISSN: 2582-4341

METHOD

Fig. 1: Overview of the model 6.

AIMS & OBJECTIVE along with monitoring Airport premises

A machine learning framework was and conserving Energy.
developed to predict the quarterly average
airfare price at the market segment level. This web App is cross-language compiled
Several features were extracted from the software [11]. Its machine learning
datasets and combined with algorithm and API are written in Python
macroeconomic data, to model the air Language and its various Frameworks and
travel market segments. With the help of Libraries.
the feature assortment methods, our
proposed model can forecast the quarterly These days, Python is covering almost all
average airfare price with an adjustment. aspects of the IT industry Software
The cheapest available ticket for a given production. HTML, CSS, and JavaScript
date gets more or less expensive over time. have been used for the front-end UI
interface and back-end connectivity has
APPLICATIONS been handled by PHP language. Flask has
RavelSaathi is an AI-enabled Web been used for API building, and Wamp
Platform-based application. It serves as the Server to host WEB APP on local
day-to-day companion for Airport windows PC.
Authority individuals. It can be useful for
both aviation professionals and their FUTURE SCOPE
customers.  More routes can be added and the
same analysis can be expanded to
They can intelligently set the Fare price of major airports and travel routes in
flight tickets, can compare with India.
competitive industries, and can also  The analysis can be done by increasing
provide environmentally friendly and the data points and increasing the
energy-saving solutions. More historical data used that will train the
specifically[8], this system is designed to model better giving better accuracies
allow the Aviation industry to set and more savings.
competitive fare prices for their Flight  More rules can be added to the Rule-
service and obtain beneficiary margins; based learning based on our

HBRP Publication Page 1-5 2022. All Rights Reserved Page 3

 Advancement and Research in Instrumentation Engineering
Volume 5 Issue 1
e-ISSN: 2582-4341

understanding of the industry also Intelligent Engineering (pp. 191-203).

incorporating the offer periods given Springer, Singapore.
by the airline. 4. Wang, T., Pouyanfar, S., Tian, H.,
 Developing a more user-friendly Tao, Y., Alonso, M., Luis, S., & Chen,
interface for various routes giving S. C. (2019, July). A framework for
more flexibility to the users. airfare price prediction: a machine
learning approach. In 2019 IEEE 20th
CONCLUSION international conference on
From our thorough analysis of each of the information reuse and integration for
18 routes, we can define the following. data science (IRI) (pp. 200-207).
Flight prices almost always remain IEEE.
constant or increase between the major 5. Papakostas, G. A., Diamantaras, K. I.,
cities. & Papadimitriou, T. (2017). Parallel
pattern classification utilizing GPU-
Tourist routes that offer services involving based kernelized Slackmin algorithm.
Tier-2 cities of the country have uneven Journal of Parallel and Distributed
treads related to the increase and decrease Computing, 99, 90-99.
of airline ticket prices. Routes with data 6. Francis, G., Fidato, A., & Humphreys,
collected over a longer duration of time I. (2003). Airport–airline interaction:
tend to facilitate much more accurate the impact of low-cost carriers on two
predictions in the model and thus lead to European airports. Journal of Air
higher average savings. Transport Management, 9(4), 267-
273.
We were successfully able to analyze each 7. Koopmans, C., & Lieshout, R. (2016).
route and generalize the entire project Airline cost changes: To what extent
based on terms. We have also successfully are they passed through to the
busted some of the typical myths and passenger?. Journal of Air Transport
misconceptions related to the airline Management, 53, 1-11.
industry and backed them up with data and 8. Lee, S., Seo, K., & Sharma, A. (2013).
analysis. Corporate social responsibility and
firm performance in the airline
REFERENCES industry: The moderating role of oil
1. Groves, W., & Gini, M. (2011). A prices. Tourism management, 38, 20-
regression model for predicting 30.
optimal purchase timing for airline 9. Nair, V., & Hinton, G. E. (2010,
tickets. January). Rectified linear units
2. Liu, T., Cao, J., Tan, Y., & Xiao, Q. improve restricted boltzmann
(2017, December). ACER: An machines. In Icml.
adaptive context-aware ensemble 10. Breiman, L. (2001). Random forests.
regression model for airfare price Machine learning, 45(1), 5-32.
prediction. In 2017 International 11. Hong, D. H., & Hwang, C. (2003).
Conference on Progress in Support vector fuzzy regression
Informatics and Computing (PIC) (pp. machines. Fuzzy sets and systems,
312-317). IEEE. 138(2), 271-281.
3. Boruah, A., Baruah, K., Das, B., Das, 12. Gerardi, K. S., & Shapiro, A. H.
M. J., & Gohain, N. B. (2019). A (2009). Does competition reduce price
bayesian approach for flight fare dispersion? New evidence from the
prediction based on kalman filter. In airline industry. Journal of Political
Progress in Advanced Computing and Economy, 117(1), 1-37.

HBRP Publication Page 1-5 2022. All Rights Reserved Page 4

 Advancement and Research in Instrumentation Engineering
Volume 5 Issue 1
e-ISSN: 2582-4341

13. Guo, X., Zhao, Q., Zheng, D., Ning,

Y., & Gao, Y. (2020). A short-term
14. Load forecasting model of multi-scale
CNN-LSTM hybrid neural network
considering the real-time electricity
price. Energy Reports, 6, 1046-1053.

HBRP Publication Page 1-5 2022. All Rights Reserved Page 5