Bus ticket price monitoring, with 42% more tickets sold within four months of launch

Competitor prices from bus carriers and ticket aggregators, collected on a schedule and fed back into our client's own pricing

Data integration and warehousing

Competitor bus ticket prices compared in a table on a laptop: the company's own price beside two carriers and two aggregators for each route and departure

Who we worked with

Bus travel company that wanted competitors' prices

Our client is a bus travel company. On a ticket aggregator, buyers see several carriers' prices for the same trip in one list, and a price set without knowing the others is a guess. So our client wanted to see what competitors charged, with prices from both bus carriers and ticket aggregators, collected to a specification it set itself.

Client under NDA

What we did

How competitor prices reach our client's pricing

Scheduled collection on a multi-region spot fleet

Our client's main application starts the data collection on a schedule, and Apache Airflow runs the Python workflows that hand the requests out to a fleet of spot instances in several AWS regions. Prices from carriers and aggregators land in file storage, and PostgreSQL holds the collected prices that go back into the main application, where our client adjusts its own. Spot capacity is cheap and can be taken back at short notice, so a request cut off when an instance is reclaimed is retried.

Weekly at launch, daily later

At launch the workflows ran once a week. As the system evolved they ran daily, so our client worked from fresher competitor prices and could adjust its own more often. Weekly is a sensible place to start. A team learns what it does with a competitor's price before it pays to collect one every day.

Room for the next aggregator

We designed collection so a new aggregator can be added later, on its own or in parallel with the others. Each source has its own Apache Airflow workflow, so a new one is one more workflow, and the existing ones stay untouched.

Diagram: one Airflow workflow per source sends requests to a multi-region spot fleet; prices return to the application

After launch

42%

More tickets in 4 months, per our client

Daily

Competitor price collection

Need competitor prices inside your own system?

Tell us which competitors and sources you need prices from, how often, and which system should use them. We reply within 24 hours.

Case studies

More data projects from our team