A Curseword-Based Scale for Temperature

5,400 geolocated tweets with phrases like "it's hot as hell" or "it's cold as a bitch" were collected from August 2017 to January 2018, and the outdoor temperature collected for each.

The result is a derived curseword-based temperature scale: "it's hot as hell" is 86°F, statistically speaking.

By @jimwebb, and presented at Hack and Tell DC (Presentation: Keynote | PDF). Thanks to @metasemantic for inspiration and code.

Observed median temperature, by phrase (n>10)

Most popular phrases and median temperatures (°F)

Phrase	Count	Median Temp (°F)
(it's) cold as fuck	623	41°F
cold as shit	497	45°F
cold as hell	466	48°F
cold as balls	78	37°F
cold as a bitch	43	37°F
---------------------------	-------	---------------
hot as hell	549	86°F
hot as fuck	495	86°F
hot as shit	261	86°F
hot as balls	197	84°F
hotter than hell	58	85°F

Detail: hot subjects (n>5)

Displayed with frequency, median (black) and 95% confidence (gray)

Detail: cold subjects (n>5)

Displayed with frequency, median (black) and 95% confidence (gray)

"Hell" can be hot or cold

"Cold as hell" (48°F) and "hot as hell" (86°F) exist together, and hell isn't the only subject with this duality:

Phrase	When Hot (°F)	When Cold (°F)
dick	92	36
satan's balls	90	34
a witch's tit	90	26
a bitch	89	38
a mf	88	42
a mother	88	29
ever	87	59
f	87	45
tits	85	40
shit	85	45
hell	85	47
fuck	84	41
balls	84	36
heck	78	28

"Boogers" are always cold; "The Devil's Dick" is always hot

Some subjects are used in one context (either hot or cold), but not both. "Boogers" and "Mars" are always cold; "the devil's dick" and "two rats (fucking)" are always hot:

Data set

Dataset provided in data/collected-tweets.csv; tweets with temperatures added are a SQLite database, in data/collected-tweets.db. Contact me for raw tweets (250MB).

Data collection

python src/0-scrape.py

Requires an empty folder in the local directory named raw_firehose and a file named access_tokens.json with your Twitter API keys:

{
    "key":"XXXXXXXXXXXXXXXXXXXXXXXXX",
    "secret":"XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX",
    "access":"XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX",
    "access_secret":"XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
}

Processing

Set up SQLite database for processing (one-time)

php src/1-dbsetup.php

Parse the tweets and organize by phrasing

php src/2-collect-tweets.php

Poll the Dark Sky API for temperatures (edit file with your Dark Sky API key)

php src/3-get-weather.php