Thanks for reading. Yes, absolutely agree. I created a separate view in Analytics to record the null conversion sessions and export the journeys from there, removing those with any ‘proper’ conversions and adding those to the model. I’ve had this sitting as a first paragraph in the draft of part 2 of this article, really do need to get back to it and finish it off!
Thanks for this, great read! For those that use Ableton Live, there is a Ganz Graf package for Max4Live to produce some visuals on the fly to your music. Also, just wanted to take the opportunity to mention Aphex Twin’s Equation video! https://www.youtube.com/watch?v=M9xMuPWAZW8
Thanks for reading, and a good point — it doesn’t really add any value at all, other than perhaps catching a bit more attention if it were posted on social media. I had tried to convey this in the sub-heading and text of the article, but perhaps didn’t make it quite clear enough. This was just an exercise to spend a bit more time with ggplot2…
Hi Yuki, thanks for reading. Have you installed the RGA package and then loaded it with
library(RGA)? If you’re using R Studio, click on the ‘Packages’ tab of the Files / Plots / Packages… pane and RGA should be there with tick mark in the box; alternatively, run
sessionInfo() in the console and you should see
RGA listed as an attached package. Hopefully that’s all the problem is.
Nice read, and great point about putting the accuracy score of your model into context of how balanced the dataset is. Using the
confusionMatrix() function from the
caret package in R is quite useful for that, as you can easily compare the accuracy to the no information rate and very quickly get a feel for how much value your model is adding compared…
Interesting article. I agree with your point about asking for a PhD and experience is a way to try and people who have been pre-qualified, and think there’s definitely something to be said for hiring people for what they can do, not just on what they’ve done in the past. This isn’t exclusive to data science of course, but it seems as though a lot of…
Some very interesting points, particularly about the increasing use of influencer marketing as part of the modern marketing mix. I just had a few questions about your article:
Really nice bit of work, and great use of the Spotify and Genius APIs. Having also been a Taylor fan since the first album, I had started off a similar project that I meant to get finished before the Reputation concert, but never quite got around to finishing it; between your post and this one, I think I’ll just leave it!
Really nice article. Coming from > 10 years of academic science into marketing, the research methodology and statistical rigour is something I absolutely believe in, even outside of academia. However, I would say that there are different interpretations of the word ‘story’.