Taking a stroll up Ballmer’s Peak

the perfect evening counter-balance to that nice, hot mug ofshut the f* up you poured yourself earlier in the morning

There’s this awesome micro-brewery called Whiprsnapr operating in my neck of the woods. In 2019, they had this Unfiltrd project where they brewed one special edition beer every week. It was magnificent.

As the year waned and the heap of cans waxed, I thought it’d be fun to have a dashboard showing me my Unfiltrd scorecard; which weekly beers I prefered, which ones I missed, that kind of stuff.

So I cracked myself a fresh can of inspiration, and dove in.

Brewing the data

In our modern age epicurian pleasures are not complete without a side help of social sharing. And indeed, the beer world has Untappd.

Has luck would have it, Untappd offers not one but two API to access its alcoholized knowledge base. One of them, Uptappd for business, is graphql-based and would have been perfect for my needs but, alas, business I ain’t. The second, a more classical REST interface, is however open to anybody who asks nicely.

Tantalizingly enough, the brewery endpoint of that REST interface can provide part of its beer list, but doesn’t provide any obviously documented pagination method. Huge bummer. So I had to resort to a more generic search for the brewery name, at the risk of getting a few false positives along the main catch:

// get_whiprsnpr_beers.js

const axios = require('axios');
const { writeJson } = require('fs-extra');

const base_url = "https://api.untappd.com";
const access_token = "DEADBEEF";

fetch_all_beers().then(
  beers => writeJson( 'beers.json', beers )
);

async function fetch_all_beers() {
  const beers = [];

  let offset = 0;

  while(true) {
    const res = await axios.get( `${base_url}/v4/search/beer`, {
      params: {
        access_token,
        q: 'whiprsnapr',
        offset:  25 * (offset),
      }
    });

    offset++;

    const more_beers = res.data.response.beers.items;
    beers.push( ...more_beers );
    if( more_beers.length == 0 ) return beers;
  }
}

That returns 156 different beers. Nice! Ah, but the search doesn’t return all the data associated with a beer. So I have to return for a second, beer-by-beer pass.

# get_beers_data.pl

use 5.30.0;

use File::Serialize;
use DDP;
use DateTime::Format::Flexible;
use LWP::Simple;

my @beers = map { $_->{beer} } grep {
    $_->{brewery}{brewery_id} == 55938
} @{ deserialize_file 'beers.json' }@beers;

$_->{created_at} = 
  DateTime::Format::Flexible->parse_datetime( $_->{created_at} )
    for @beers;

# only interested in 2019
@beers = grep { $_->{created_at}->year == 2019 } @beers;


for ( @beers ) {
  sleep 3;
  say $_->{beer_name};
  my $id = $_->{bid};
  mirror(
    "https://api.untappd.com/v4/beer/info/$id"
      ."?access_token=DEADBEEF&compact=true", 
    "beers/$id.json"
  );
}

There, the data of 55 different beers have been retrieved. A typical one looks like:

"beer": {
    "bid": 3030726,
    "beer_name": "Bells Corners Beast",
    "beer_label": "https://untappd.akamaized.net/site/assets/images/temp/badge-beer-default.png",
    "beer_label_hd": "",
    "beer_abv": 6.66,
    "beer_ibu": 66,
    "beer_description": "SMaSH IPA using Canada's first born and bred hop - Sasquatch.  3lbs hops/barrel!",
    "beer_style": "IPA - American",
    "is_in_production": 1,
    "beer_slug": "whiprsnapr-brewing-co-bells-corners-beast",
    "is_homebrew": 0,
    "created_at": "Fri, 11 Jan 2019 01:44:48 +0000",
    "rating_count": 22,
    "rating_score": 3.86364,
    "stats": {
    "total_count": 29,
    "monthly_count": 0,
    "total_user_count": 22,
    "user_count": 0
    },
}

Informative, but not pretty. Let’s address that.

Pouring the stats

As I’m currently cutting my teeth on all that is React, why not display our suddy data as a semi-mobile-friendly app? Overkill? probably. But educationally so.

I won’t bore you with the whole application. Suffice to say that it’s pretty barebone React using Material UI as its component library (it has decent documentation, is fairly sane, and provides a decent amount of components), and uses Chart.js for all the graphy stuff.

The result so far? A wee little app that has two screens. The first displays all the Whiprsnapr beers of 2019 (according to what was retrieved from Untappd. A quick perusal makes me suspect a few beers slipped through the net). If I was to complete the Untappd integration, it’d be fairly easy to tailor-suit the list according to the visitor’s Untappd profile.

The second page is a gleeful dump of graphs. Like, if there was any doubt the crew like their IPAs and stouts, it’s now scientifically proven:

Which beers won the popularity contest? Wonder no more:

Okay. Rating is one thing, but do people actually put their mouth where their beer belly is?

(that ratings vs check-ins graph is in need of a little grooming. The beers scoring a zero (unthinkable!) are probably some that didn’t get any rating. And the two beers that have an insane amount of checkins are the Risqué, not a weekly edition but a regular that appeared during the year, and the Jollypop! that looks like it’s aggregating the 4 variations that were made on successive weeks.)

And just for giggles: how about a plot of abv versus ibu?

For the road

I could play much more with that app. Untappd integration, more graphs, having the existing graphs reveal more information (like, the abv vs ibu graph is nice, but it could speak even more if the points were color-coded by beer style), etc. But for a first pass, it’ll do.

Mind you, I’m itching for some R fun, so there might be a second pass soon…