Statistical analyses were performed on the continuous and categorical variables against the markup on minimum ticket prices. I found it interesting that markups had a slight, but significant negative correlation to Spotify popularity and followers. Day of the week doesn't seem to be as important as I would have believed, as the analysis did not find weekend events to have markups significantly higher than the mean or weekdays. I also found it to be quite interesting how a small promoter based in Atlanta (Masquerade), generated much higher markups than any of its competitors. Events self-promoted by the venue also commanded higher markups than events from big names like Live Nation and House of Blues. There also doesn't seem to be an obvious connection between the state in which the venue is located, and overall markup. New York and Nevada have the most events, but are on opposite ends of the spectrum, while Louisiana and Georgia command higher markups than places with high cost of living such as California.
- Ticketmaster Minimum Price
- Spotify Popularity
- Spotify Followers
- Quantity of Tickets Listed
- Length of Presale (Days)
- Number of Days on sale
- Days until Show
- Artist count per event
Because markups followed a logarithmic distribution, both Spearman and Pearson correlations were calculated.
Column | PearsonR | PearsonR_pvalue | Pearson_NullHypothesis | SpearmanR | SpearmanR_pvalue | Spearman_NullHypothesis |
---|---|---|---|---|---|---|
TM_min | 0.011576 | 4.781360e-01 | Accept | -0.438978 | 8.478316e-177 | Reject |
avg_ticket_listings | -0.09455 | 0.000000 | Reject | -0.19922 | 0.000000 | Reject |
spotify_avg_followers | 0.04988 | 0.002227 | Reject | -0.09213 | 0.000000 | Reject |
spotify_avg_popularity | 0.04966 | 0.002328 | Reject | -0.09179 | 0.000000 | Reject |
presale_length | -0.07767 | 0.000002 | Reject | -0.18001 | 0.000000 | Reject |
days_on_sale | -0.04559 | 0.005186 | Reject | -0.16739 | 0.000000 | Reject |
days_until_show | -0.07371 | 0.000006 | Reject | 0.01310 | 0.422253 | Accept |
artist_count | 0.01077 | 0.509459 | Accept | 0.01036 | 0.525406 | Accept |
Spearman R values are generally higher than their Pearson R counterparts. TM minimum price, Artist Count and potentially Days Until Show are not correlated to price markups and may not be relevant features.
- Genre
- Subgenre
- Day of Week
- Promoter
- Ticket Source
- Venue State
T tests were used to compare the markups of each feature category to the whole dataset's markup mean. Then an ANOVA F-test was used to determine if the means of categorical features were significantly different. After confirming a statistically significant difference in markup means amongst categories, a Paired Tukey test was used to compare every unique pair of categories. Because categories had unequal sample sizes and unequal variances, the ANOVA F-Test and subsequent Paired Tukey test may not be the most ideal choices. However, they were used for the sake of simplicity since the final product of this project is not a pairwise analysis of categories.
The results corroborated the conclusions drawn from the initial graphic data exploration.
The Dance/Electronic, Jazz, Pop, R&B, Religous, Rock, Undefined, and World genres had markups significantly different from the mean.
Pop & Country, Pop & Rock, Rock & Undefined, Rock & World pairs had significantly different means.
Adult Contemporary, Club Dance, Gospel, Jazz, Latin, Pop, R&B, Undefined, and World subgenres had significantly different markups from the overall mean.
The Latin & Pop and Pop & Soul pairs had significantly different markups.
Markups for events on Sunday and Monday were significantly below the mean markup.
No statistically significant difference was found for markups between days of the week.
Most promoter categories have statistically significant differences from the overall mean. The null hypothesis is rejected for AEG Live, Crossroad Presents, Live Nation, and 'Other' for having markups significantly below the mean, and Masquerade, and 'Promoted by Venue' having markups significantly greater than the mean.
Events by Masquerade had markups significantly higher than all other promoters
Events with tickets only on Stubhub have significantly higher markups than those of the dataset. Events with tickets on both Stubhub and SeatGeek have significantly lower markups.
Differences are statistically significant for tickets on both platforms vs. Stubhub, and SeatGeek vs. Stubhub.
CA, FL, IL, KY, MA, MI, NC, NV, OH, PA, TX, and VA have markups significantly lower than the mean. GA, LA, and NY have markups significantly higher than the mean
CA & NY, FL & NY, GA & NV, IL & OH, IN & NY, KY & NY MA & NY, ME & NY, MI & NY, MO & NV, NC & NY, NJ & NY, NV & NY, NY & OH, NY & PA, NY & TX, NY & VA have statistically different means. New York has significantly higher markups than most other states, while Nevada has significantly lower markups than most states.