While browsing the San Francisco subreddit, I found a data set compiled by Max Woolf that included a list, maps, and charts of San Francisco area bars by price, happy hours, and ratings. Max Woolf retrieved the data from Foursquare using a couple of Python scripts, and output the data into a table sorted by price, happy hour and rating. Since this combines a couple of my favorite things – data analysis and beer – I thought this would be a fun bit of data to plug into Chartio - read on to see what I discovered from the data!
One of the first questions I wondered was, “What are the best rated bars?” Perhaps unsurprisingly, most people simply like bars. There was little variance in how much people liked a particular bar, and most did not rate poorly. The worst average rating was a 5/10, the average was 8/10 and the highest was 9.6/10. The average rating per bar had a standard deviation of almost exactly 1, so really very little distribution across the ratings.
It turns out that looking at ratings by price doesn’t make much of a difference either. There was a slightly higher rating for the more expensive bars. Either San Franciscans don’t mind spending a little extra money when it comes to cocktails, or they’re more forgiving when they’re flowing cheaply.
My last try to see some variance with ratings, was examining the most rated bars in the Bay Area - apparently Zeitgeist is the most rated bar. They have 5204 ratings, and their rating on FourSquare is 9.37. Zeitgeist does not have the highest rating, as you can see, but it appears that the most rated bars have pretty good ratings.
Next I moved on to the different types of bars, because you know that San Francisco has a plethora of options. I discovered that the most options, a total of 416 bars, are available at a moderate price range ($$) only includes 6 different bars, ranging from American, Cocktail, New American and a Japanese Restaurant.
And finally, the most important question - who has happy hour?? About half the bars - 306 different places - have happy hour.
Let’s see if we can increase the sample size for our next analysis. Cheers!