How to buy a ticket as cheap as possible or strike monitoring on dynamic pricing

How is the most profitable to buy a ticket?

Any more or less advanced web user knows such options as
')

• buy in advance
• look for routes with transfers
• hidden-city ticketing
• monitor charter flights
• search incognito browser
• use airline mile cards, all sorts of bonuses and promo codes

The full list of lifehacks somehow did Tinkoff-Journal, I will not repeat

Now answer the question - how often did you find yourself in a situation when you bought a ticket, and then it became cheaper?

I got and was a little hurt. This happens most often in the summer season because of the thing called dynamic pricing.

Here is a real schedule of price changes for today's flight A4 203 Rostov-on-Don - St. Petersburg of Azimut Airlines. On the x-axis - hours before departure, on the y-axis - ticket price



The schedule shows that 20 hours before departure you can buy a ticket at the lowest possible price - 4090 rubles. At the same time for 72 hours a ticket costs more than 2 times more expensive - 9390 rubles. The schedule received stupidly by parsing on cron every 15 minutes, entering the results into the database and visualizing the data via Chart.js. Who cares - this is proof. Now there is data only on flights between Rostov and St. Petersburg, but it’s not a problem to fasten other cities of the route network.

Such price fluctuations, as I understand it, are caused by the fact that the dynamic pricing algorithm by sales dynamics feels that not all tickets can be sold out and reduces the price, guided by the logic “it is better to sell the remaining tickets a little cheaper than to leave empty seats”. In other words, the higher the demand and the fewer the seats, the higher the ticket price .

Analysis of 84 flights between Rostov and St. Petersburg gave just such a picture (on the x axis - days before departure, on the y axis - ticket price)



From it we see that the really best saving strategy is to buy tickets in advance (starting from the 80th day before the trip, the price starts to grow). However, we also see from here that if, say, we have 30 days left before the trip, then it’s best not to hurry and wait a bit - there is a chance that the price will drop from 9,100 rubles to 6,100 and you will save 3,000 rubles. And given the information from the example above, it is likely that 20 hours before departure, the price may again be as low as possible.

In connection with the above, I have the following questions to the habrasoobschestvu

1) Questions for those who work in the industry .
Is the price dynamics of other airlines similar or is it a special case of Azimuth Airlines?
What factors influence pricing a priori? The number of days before departure, the day of the week (holidays or school holidays), time of year, time of day, what else?

2) Questions for representatives of agents (Aviasales, Skyscanner, OneTwoTrip Yandex.Flights, Tinkoff.Travel, etc.).
Do you collect data on price movements? If so, what do you need to access this data? Are there any partner APIs already, if not, can you unload the database?

3) Question for all who fly .
What services of cheap air ticket notifications in the direction of interest do you use? Do you need a service of price drop forecasts in a direction of interest to a selected date range?

Personally, I use the Aviasales subscription which looks like this:



She has two significant drawbacks:

1. Low responsiveness. Notifications can only come to email. Personally, I rarely look through mail, I would prefer Telegram bot
2. No prediction. Personally, before I subscribe, I would like to see what the probability is that the price will drop based on historical data.

In addition, letters come out of the ordinary. Now, it seems, Aviasales subscription does not work at all - a link to confirm a new subscription does not come.

There are also subscriptions Yandex.Flights and tutu.ru but, as I understand it, they allow you to track price changes only on a specific date.

Plus it is not clear at all how often all these services check prices - once a minute, hour, day?

PS: Infa, by the way, is relevant not only for airlines, but also for traveling by train. Online Railways straight there is an article about dynamic pricing .

PPS: What else can you read the topic
https://habr.com/en/company/iqplanner/blog/297540/
https://habr.com/en/company/friifond/blog/291032/