WeRateDogs is a Twitter account that provides mostly humourous ratings for dogs from images and descriptions provided by the dog's owners. This project involved gathering data about the tweets, assessing the quality and tidiness of the data, cleaning the data, and then performing exploratory data analyses on the data.
NOTE: This exploration was not intended to imply causation between any of the examined metrics. Any trends shown here are at best correlation rather than causation.
The data wrangling steps are documented in the accompanying Jupyter notebook.
To deal with some tweets with ratings not denominated by 10, I normalized the ratings to a 10-point scale to make comparisons and calculations more intuitive. From this normalized rating, the average rating was 10.88/10
. As the WeRateDogs account once famously proclaimed, they're indeed good dogs!
The most retweeted tweet got nearly 80k retweets, which is quite impressive for a tweet posted back in 2016. The tweet text and image are show in the picture below.
It's a very cute dog indeed!
The most favorited tweet was favorited more than 132k times, which is also very impressive for a tweet posted back in 2016. The tweet text and image are show in the picture below.
Also a very cute dog!
The bar chart below shows the average rating for each dog stage.
To summarize the dog stages briefly (more details available in the WeRateDogs book on Amazon):
doggo
: a big pupper, usually olderpupper
: a small dog, usually youngerpuppo
: a transitional phase between doggo and pupper, sometimes viewed as the dog equivalent of a teenagerfloofer
: any dog that has a lot of furLooks like furry dogs performed the best. 🐕
This metric was computed only for the names that appeared more than 3 times in the dataset. The bar chart below shows the top 5 highest rated dog names.
This dataset indicates that "Sophie" is a pretty good name for a dog! 🐶
Here's one of the tweets featuring a dog named Sophie:
The bar charts below show the average counts of retweets and favorites for each dog stage.
It seems the elder dogs were the most shared whereas the dog teenagers were the most loved. Interesting!
As with the highest rated dog names, this metric was computed only for the names that appeared more than 3 times in the dataset.
The bar charts below show the average counts of retweets and favorites for the top 5 most retweeted and favorited dog names, respectively.
"Bo" appears to be both the most shared name and the most loved, impressive! Also the top 5 names are the same for both retweets and favorites, so it seems that the most popular names are also the most loved. All 5 are clearly nice names for dogs. 👌🏾🐶
Here are tweets for each of these names:
Can you blame them for ranking so highly? 😍🤷🏾♂️
From the analyses and visualizations above, the following insights were found:
80k
retweets, whereas the most favorited tweet was favorited more than 132k
times12.0
out of 10
15
out of 10
It must be restated that this exploration was not intended to imply causation between any of the examined metrics. Any trends shown here are at best correlation rather than causation.