Predicting ticket sales for events

[cummins070]

Hi there,

A few questions I would really appreciate some guidance on!

I am trying to forecast ticket sales for conferences. We run a major conference each year. I have daily ticket sales for 2 previous years. We sell tickets 1 year in advance, e.g. once the conference is over for 2021, ticket sales start for 2022 the next day.

Typically we see a exponential increase in ticket sales as we get closer to the event (people like to leave it to the last minute to buy tickets!)

We have a few major factors that influence sales:

• Major announcements e.g. when we announce the date and venue
• The price of the tickets, we start off with early bird pricing and prices gradually increase every few months until the date of the event
• Marketing efforts e.g. emails sent out with reminders of prices increase in X number of days
• Paid Digital ads e.g. daily ad spend from Google / Facebook etc.

I'm wondering how best to set this up with neural prophet? I'm assuming I would want to have the following in my dataframe:

• Date (daily)
• Ticket sales
• The event label e.g. "1" could be for the 2021 event, "2" for 20202?
  -The ticket price on the given date
• Any major announcements (assuming I would add these as special events)
• Dates of emails sent announcing ticket price increases
• Daily ad spend from the marketing platforms
• The date of the event each year? again using special events

Is there anything else I should be considering? I've used prophet previously, but on a much simpler forecasting task and I'm certainly still very green. Just trying to figure out the best approach and factoring in as many major variables as possible!

Many thanks :)