A few months ago I posted a few interesting bits of research on data from SEDE. It was fairly well-received, so I figured I'd do some more analysis. I might make this a weekly thing.
I decided to start with tags. This post will show some data relating to the most used tags, the popularity of posts based on the tags they use, and the most common combinations of tags.
All of this data is from April 6, 2020.
Most used
- code-golf: 9540 (84.16%)
- string: 1750 (15.44%)
- math: 1362 (12.01%)
- number: 1087 (9.59%)
- ascii-art: 1069 (9.43%)
- code-challenge: 810 (7.15%)
- sequence: 680 (6.00%)
- kolmogorov-complexity: 564 (4.98%)
- array-manipulation: 530 (4.68%)
- arithmetic: 480 (4.23%)
- decision-problem: 474 (4.18%)
- popularity-contest: 395 (3.48%)
- number-theory: 354 (3.12%)
- integer: 307 (2.71%)
- graphical-output: 303 (2.67%)
- restricted-source: 290 (2.56%)
- geometry: 287 (2.53%)
- random: 278 (2.45%)
- combinatorics: 274 (2.42%)
- primes: 272 (2.40%)
Highest average score
Note: Only contains tags with at least 10 questions
- tetris: +83.0 (15647 views)
- game-of-life: +82.9 (16908 views)
- self-referential: +59.9 (8444 views)
- new-years: +50.0 (13467 views)
- color: +49.1 (10966 views)
- cellular-automata: +48.9 (7895 views)
- code-trolling: +48.0 (53586 views)
- underhanded: +44.3 (11880 views)
- answer-chaining: +41.5 (7826 views)
- error-message: +39.9 (11323 views)
- hello-world: +39.8 (10270 views)
- stateful: +38.7 (9483 views)
- busy-beaver: +36.3 (6072 views)
- obfuscation: +31.8 (7838 views)
- factorial: +31.5 (5103 views)
- language-design: +31.2 (1907 views)
- popularity-contest: +31.0 (8131 views)
- radiation-hardening: +30.7 (2908 views)
- interactive: +29.9 (4293 views)
- king-of-the-hill: +29.8 (3398 views)
Most common combinations
- code-golfstring: 1658
- code-golfmath: 1214
- code-golfascii-art: 1024
- code-golfnumber: 1008
- code-golfsequence: 625
- code-golfkolmogorov-complexity: 527
- code-golfarray-manipulation: 508
- code-golfdecision-problem: 457
- code-golfarithmetic: 447
- code-golfnumber-theory: 328
- code-golfinteger: 288
- code-golfgeometry: 262
- code-golfrandom: 254
- mathnumber: 248
- code-golftips: 243
- code-golfprimes: 232
- code-golfdate: 229
- code-golfgraphical-output: 227
- code-golfcombinatorics: 224
- code-golfgrid: 220
Most common combinations (compared to individual usage)
Note: Only contains tags with at least 10 questions
- fewest-operationsgolf-cpu: 5 (33.33%, 100.00%)
- cellular-automatagame-of-life: 12 (30.77%, 46.15%)
- internetstack-exchange-api: 19 (25.00%, 57.58%)
- assemblymachine-code: 5 (23.81%, 62.50%)
- cc++: 13 (24.53%, 39.39%)
- code-golfstring: 1658 (17.38%, 94.74%)
- programming-puzzlec#: 9 (22.50%, 37.50%)
- functional-programminghaskell: 5 (21.74%, 33.33%)
- king-of-the-hilljava: 24 (16.44%, 54.55%)
- matrixlinear-algebra: 30 (15.15%, 73.17%)
- graphical-outputimage-processing: 50 (16.50%, 40.65%)
- number-theoryprimes: 73 (20.62%, 26.84%)
- atomic-code-golflogic-gates: 11 (16.92%, 36.67%)
- numbersequence: 202 (18.58%, 29.71%)
- code-golfmath: 1214 (12.73%, 89.13%)
- error-correctioncoding-theory: 2 (18.18%, 28.57%)
- rubiks-cubecubically: 3 (12.50%, 75.00%)
- binary-treetree-traversal: 4 (19.05%, 23.53%)
- tipspython: 39 (14.39%, 37.14%)
- popularity-contestunderhanded: 46 (11.65%, 92.00%)
Interesting Takeaways
- Code golf makes up over 84% of the challenges on this site
- Despite now being strongly discouraged, popularity-contest, underhanded, and code-trolling have a surprisingly high average score
- 92% of all underhanded challenges were popularity contests
- A few popular questions, such as this one, heavily influence the average score of some tags like tetris
- Of all the top 20 most often combined tags, math and number are the only ones to not include code-golf
- Just 11 out of 277 tags have not been used in combination with code-golf
Just 11 out of 277 tags
? \$\endgroup\$