[33 / 2 / 27]
Quoted By: >>57357988 >>57358069 >>57358094 >>57358130 >>57358231 >>57358363 >>57358428 >>57360597 >>57362651 >>57362776 >>57364324 >>57365355 >>57366603 >>57369402 >>57370728 >>57372707 >>57372972
Premise: I wanted to know how Gura affected Advent's numbers. My thought was that if Gura's overlapping did negatively affect Advent's numbers then Advent should have higher numbers when Gura is not streaming and lower numbers when she is. Additionally, once Gura stopped streaming then Advent should receive a boost in numbers, however small, since the viewers who Gura stole would watch Advent instead.
Methodology: I took average CCV from July 31 to August 28 and shoved them into Jupyter Notebook. For calculating Advent's average CCV, I averaged the CCV of the ones who actually streamed that day. The blue bars are when Gura overlaps and the red bars are when she doesn't. If a member streamed more than once during the afternoon, I took the maximum. I excluded Fuwamoco mornings since I don't think they're applicable to this analysis. I also excluded the first collab on July 31st since it was such a big outlier and they were all on one channel.
Results: For the 18 days that Gura was streaming, the Advent average with Gura was 11986.9 and without Gura was 12291.875. However, this includes a pretty big outlier of the monetization stream, where Advent all got pretty high numbers. If we exclude the outlier, with Gura is 11985.5 and without is 10638.285.
The average advent views while Gura was streaming was 12121.89 and 8011.09 after Gura stopped streaming. If we limit it to 4 days before Gura stopped streaming and 4 days after Gura stopped streaming, it is 8882.75 before and 7989.5 after.
Conclusion: While I was gathering the numbers I had to look at a lot of thumbnails, which made me realize that all the Advent girls and Gura are super cute, which made me happy. My secondary conclusion is that pretty much all of this data is junk since there is so much variation in such a limited dataset. View count swung wildly depending on if they were doing karaoke, superchat readings, streamer-unfriendly games, etc., and with only 18 days of data where Gura was streaming the data is very untrustworthy. Also there was the whole Shiori taking a 10 day break thing, which affected the data in unknown ways.
tl;dr no conclusion can be made because the data is both small and full of outliers, just enjoy the girls because they're cute.
Methodology: I took average CCV from July 31 to August 28 and shoved them into Jupyter Notebook. For calculating Advent's average CCV, I averaged the CCV of the ones who actually streamed that day. The blue bars are when Gura overlaps and the red bars are when she doesn't. If a member streamed more than once during the afternoon, I took the maximum. I excluded Fuwamoco mornings since I don't think they're applicable to this analysis. I also excluded the first collab on July 31st since it was such a big outlier and they were all on one channel.
Results: For the 18 days that Gura was streaming, the Advent average with Gura was 11986.9 and without Gura was 12291.875. However, this includes a pretty big outlier of the monetization stream, where Advent all got pretty high numbers. If we exclude the outlier, with Gura is 11985.5 and without is 10638.285.
The average advent views while Gura was streaming was 12121.89 and 8011.09 after Gura stopped streaming. If we limit it to 4 days before Gura stopped streaming and 4 days after Gura stopped streaming, it is 8882.75 before and 7989.5 after.
Conclusion: While I was gathering the numbers I had to look at a lot of thumbnails, which made me realize that all the Advent girls and Gura are super cute, which made me happy. My secondary conclusion is that pretty much all of this data is junk since there is so much variation in such a limited dataset. View count swung wildly depending on if they were doing karaoke, superchat readings, streamer-unfriendly games, etc., and with only 18 days of data where Gura was streaming the data is very untrustworthy. Also there was the whole Shiori taking a 10 day break thing, which affected the data in unknown ways.
tl;dr no conclusion can be made because the data is both small and full of outliers, just enjoy the girls because they're cute.
