When did “How I Met Your Mother” become less legen.. wait for it…

October 21, 2013
By

(This article was first published on Data and Analysis with R, for Work and Fun, and kindly contributed to R-bloggers)

…dary!  Or, as you’ll see below, when did it become slightly less legendary?  The analysis in this post was inspired by DiffusePrioR’s analysis of when The Simpsons became less Cromulent. When I read his post a while back, I thought it was pretty cool and told myself that I would use his method on another relevant dataset in the future.

Enter IMDB average user ratings of every episode of How I Met Your Mother (until season 9, episode 5).  Once I brought up the page showing rated episodes by date, I simply copy and pasted the table into LibreOffice Calc, saved it as a csv, and then loaded it into R.  As you can see in DiffusePrioR’s post (and also in mine), the purpose of the changepoint package is to find changepoints, or abrupt baseline shifts, in your data.

Now, some R code:

library(changepoint)
library(ggplot2)
library(zoo)

himym = read.csv("himym.csv")
mean2.pelt = cpt.mean(himym$Rating, method="PELT")
plot(mean2.pelt, type='l', cpt.col='red',xlab="Episode",
     ylab='IMDB Avg Rating', cpt.width=2, main="Changepoints in IMDB Ratings of 'How I Met Your Mother'",ylim=c(0,10), xlim=c(0,200))

Changepoints in IMDB Ratings of HIMYM
Above we see that the changepoint method I used has detected two baselines for the rating data.  One baseline has its value at 8.1, and the other has its value at 7.6.  The first thing you have to admit here is that despite a downwards shift in the average rating, How I Met Your Mother still seems to be pretty popular and held in high regard.  That being said, I do feel as though the earlier episodes were better, and apparently the masses of IMDB users rating each episode generally agree!  So, what’s the first episode that marked the abrupt downwards shift in average rating for the show?

himym[161,1:2]
    Ep.Num   Ep.Title
161    8.1 Farhampton

Ah, Farhampton, Season 8, Episode 1.  That seems like a natural breakpoint!  But the low score doesn’t even come until Season 8 Episode 2.  Let’s see what people have to say about episodes 1 and 2 on IMDB (forgive their punctuation, spelling, and grammar):

“Boring return I thought as once again the writing was horrible, the show seems to want to go on forever and also some of the actors in the show deserve better then what they are getting week to week. Even at the end it seems we finally see the mother of the show but we do not see her face.”

“The first episode however.was not funny at all, the thrill of maybe finding who is the mother made me think that it was not too late, but then again the joke is on me. This second episode (The Pre-Nup) only made me think that this show will not find a good finale, using all the clichés on the book of the sitcom. I just don’t care about who is the mother anymore, I just want a honesty last season, showing the characters as human beings that learn with their mistakes and move on with their lifes.”

So it appears that some people were expecting the pace to quicken in Season 8, but were disappointed.  That being said, if you look back at the graph at the beginning of my post, you’ll see that although Season 8 was the start of abruptly lower ratings, the average ratings seemed to be dropping very slowly from well before that point.  Given that the rate of change is so slow, perhaps detecting the “beginning of the end” using the changepoints package is not sensitive enough for our purpose.

That’s why I decided to bin all How I Met Your Mother episodes into groups of 10 successive episodes each, and then calculate the proportion of episodes in each bin which have an average rating under 8.0.  You can see the graph below:

eps.under.8 = rollapply(himym$Rating, width=10, function (x) mean(x < 8), by=10, partial=TRUE)
ep.nums = paste(((seq(1,19)-1)*10)+1, ((seq(1,19)-1)*10)+10, sep="-")
ep.nums[19] = "181-189"
ep.grouped.ratings = data.frame(ep.nums, eps.under.8)

ggplot(ep.grouped.ratings, aes(x=reorder(ep.nums, seq(1,19)), y=eps.under.8, group=1)) + geom_line(size=3, colour="dark red") + scale_y_continuous(name="Proportion of Epsidoes/10 With a Below 8.0 Avg Rating") + scale_x_discrete(name="Episode Number Grouping") + theme(axis.text.x=element_text(angle=-90, vjust=.5, colour="black", size=14), axis.text.y=element_text(colour="black", size=14), axis.title.x=element_text(size=16, face='bold'), axis.title.y=element_text(size=16, face="bold"), plot.title=element_text(size=20,face="bold")) + geom_vline(xintercept = 10,linetype='longdash') + ggtitle("How I Met Your Mother Ratings, Every 10 Episodes")

Grouped HIMYM Ratings

In the above graph, you can see that because I’ve set such a high threshold (8.0), the results become much more dramatic looking.  Until episodes 71-80, the show enjoyed a very high proportion of ratings of 8.0 or over (80% for 6 episode groups, 90% in one, and 100% in the first!).  It looks as though it wasn’t until episodes 91-100 that you can really notice that ratings were not staying so well over 8.0.  Which season and episodes were those?

himym[91:100,1:2]
    Ep.Num                           Ep.Title
91    5.30                          Robin 101
92    5.40              The Sexless Innkeeper
93    5.50                   Duel Citizenship
94    5.60                           Bagpipes
95    5.70                    The Rough Patch
96    5.80                       The Playbook
97    5.90 Slapsgiving 2: Revenge of the Slap
98    5.10                         The Window
99    5.11                Last Cigarette Ever
100   5.12                    Girls Vs. Suits

Well, now we have some evidence that quality started to go down earlier on in the show (although bear in mind we are still talking about a highly rated show here!).  Again, let’s go to IMDB comments to see what we can glean about these episodes:

(in reference to “The Sexless Inkeeper”) “So after waiting 9/10 months for Robin and Barney to finally stop looking for reasons not to admit their feelings to each other, Lily is euphoric when the two finally do concede that they love each other. And how does she celebrate this?, she invites them to an incredibly awkward, incredibly forward night that displays EVERYTHING that Barney and Robin hate about relationships and gives the couple a massive reason NOT to be together. Lily and Marshall then react, whence realising Robin and The Barnicle didn’t enjoy the evening, as if they are the couple that has been mistreated. Lily then dumps Robin and Barney without so much as a phone call and plays right into the pair’s issues of abandonment.”

(in reference to “Last Cigarette Ever”) “… “Last Cigarette Ever” is the worst one. There are many reasons why I feel this way, but the two big reasons are: 1. It made no sense. Up to this point, I believe only Robin had ever been shown smoking a cigarette. Now all of a sudden, they all are smokers. I’m not going to pretend that the writers haven’t made mistakes before, but this was a little ridiculous. 2. It felt like a half-hour Marlboro ad. I think the intent was to discourage smoking, but I personally felt like it did the opposite. Awful episode in an otherwise very good series.”

I’m finding there don’t tend to be many user reviews of episodes on IMDB, so this next one is from metacritic.com:

“How i met your mother is starting to annoy everyone as the jokes are repetitive, Ted, Marshall & Lilly have started to get boring. There were only 5-6 good episodes in the season which includes the PLAYBOOK. The story line and the characters seem to be awfully close to NBC’s ”F.R.I.E.N.D.S”, but friends steals the show with better writing and acting which HIMYM seems to lack. Only Neil’s role as Barney has helped CBS to keep this show on air for 5 seasons.”

In the end, it’s hard to attribute shifts in average rating to these few comments, but they at least they do highlight some of the same complaints I’ve had or heard about the show (although I didn’t care so much about the cigarette episode!!).  All we can say here is that the ratings started to decline around the start of season 5, and suffered a sharper (although still not grand scale) decline at the beginning of season 8.


To leave a comment for the author, please follow the link and comment on his blog: Data and Analysis with R, for Work and Fun.

R-bloggers.com offers daily e-mail updates about R news and tutorials on topics such as: visualization (ggplot2, Boxplots, maps, animation), programming (RStudio, Sweave, LaTeX, SQL, Eclipse, git, hadoop, Web Scraping) statistics (regression, PCA, time series, trading) and more...



If you got this far, why not subscribe for updates from the site? Choose your flavor: e-mail, twitter, RSS, or facebook...

Comments are closed.