Using ChatGPT to compare sales curves
Martin Gammeltoft
Helping live entertainment organisations use data to meet the targets for their events, by using the Activity Stream platform. Part of the industry since 2012. INTIX Board member since 2021.
Following on the previous article about using ChatGPT to analyze survey results, this time I'm trying to have ChatGPT help me with another typical task: Comparing ticket sales for two (similar) events.
For this, I have two datasets showing each transaction for the respective events.
Conclusion: This turned out pretty good. As always, the conversational flow proves valuable, as I got some ideas along the way.
BE WARNED! The second time I tried this, I couldn't get it to correctly map out the lead time - it was as if it had somehow lost the understanding of the event date. Ended up being incredibly frustrating and a complete waste of time.
Here is my (first and successful) conversation:
Me:
The attached file is a ticket sales report for an event that took place on May 26, 2023. The Sales Date column will show the data of each purchase. Based on this report, can you create a graph of sales up until the time of the event?
Please set the x-axis labeling so that it shows "days before event" and ends at 0 (the event date).
ChatGPT
The graph now displays the x-axis as "Days Before Event," counting down to the event date on May 26, 2023. The cumulative ticket sales are shown, with the days decreasing towards the event date.
(at this stage I realised I had a data error, as it was aggregating order size, not tickets, so the Y-axis was wrong in this one)
Me:
Great. This we'll call a "pace curve". I'll now upload a similar report, but for an event that takes place on August 16, 2024. Can you add that curve to the graph, for comparison?
领英推荐
ChatGPT:
Here is the updated graph showing the cumulative revenue for each performance, comparing the events on May 26, 2023, and August 16, 2024. The x-axis displays the number of days before the event, allowing you to see how revenue accumulated over time for each event
Me:
This is great. In the files are also price information about each ticket purchased. Can you make a similar graph that shows the total revenue for each performance?
ChatGPT:
Here is the updated graph showing the cumulative revenue for each performance, comparing the events on May 26, 2023, and August 16, 2024. The x-axis displays the number of days before the event, allowing you to see how revenue accumulated over time for each event.
Me:
Perfect. As a 3rd graph, could you provide the two graphs for avg ticket price over time, with the avg price showing from 400 to 650?
ChatGPT:
The graph now displays the average ticket price over time for both events, with the y-axis ranging from 400 to 650 DKK.
Subsequently, I asked it about the main differences between sales in terms of postcodes and ticket types, which gave some useful insights into the number of free tickets for each event.
Another useful follow-up would be to align from first sales day, if the focus is on a comparison of the on-sales.
Senior Managing Director
4 个月Martin Gammeltoft Very insightful. Thank you for sharing
Founder ?? Makes sense of the marketing chaos
4 个月Very cool. Did you use Chat GPT 3.5 or 4?
AI Consultant, Fractional Chief AI Officer & Founder @ Proddy.io | AI-driven product management
4 个月Great post Martin Gammeltoft and the biggest consideration with all of this… …is that this is the worst it will ever be! Case in point, yesterday I got access to the new Data analysis toolset within ChatGPT. It now supports Google and Microsoft cloud files natively, so I can share an office 365 excel sheet direct from one drive, the graphs are interactive, the data analysis is faster and more accurate… every few weeks these tools improve!