View on GitHub

TRUMPASAURUS

Analysis of word usage in political speeches


HOLD THE FRONT PAGE!

Actually I've made a terrible assumption in my calculations and I need to have a
think how to take the length of a speech into account.

--

Generally speaking it seems that the more a speech is refined the more coherent
it becomes and the unique word ratio decreases. However, this doesn't continue
indefinitely and the further you finesse the more tiresome it becomes.

Unique word ratio and average sentence length appear to be a similarly good
measure of rant or not. But there are some exceptions - like the Queen's speech
- so we must still take into account another factor. Initial banding using only
unique word ratio:

25% - repetitive, tiresome
29% - overly finessed, some repetition
32% - finessed
35% - practiced but loose
40% - quick fire, less structure
50% - rant

Exceptions:
At the Grand Hotel Thatcher spoke without an autocue yet the unique word ratio
suggests it was very well rehearsed (32%). May's Brexit speech and the 2017
Conservative manifesto also have similar low counts (29-31%) - perhaps penned by
the same hand - but surprisingly Blair's apparently quick fire response to
Paxman has a very low ratio (27%).

Theresa in Birmingham, the SNP manifesto and the LibDem manifesto boast the
lowest counts measured (22,23,24%).

----

MENTIONS IN THE MANIFESTOS ("|" means or)

Conserv	DUP	SinnFei	Labour	Green	Libdem	SNP	UKIP	
 	 	 	1	 	 	 	 	abortion
 	 	 	1	 	 	 	 	badger
 	 	 	 	 	 	 	6	blair
8	1	7	25	2	23	41	42	brexit
 	1	 	 	 	 	 	1	cameron
 	 	 	 	 	 	 	 	clegg
 	 	 	14	6	12	16	5	climate
28	6	4	66	11	68	28	21	community|communities
41	 	 	68	 	32	1	17	conservative
 	 	 	1	 	3	 	1	corbyn
13	2	1	23	 	15	5	25	crime
8	6	 	4	 	2	1	1	cyber
2	2	 	7	1	4	5	5	debt
3	13	 	16	1	8	11	24	defence
5	 	2	9	4	2	19	26	disabled|disability
29	22	4	86	19	92	23	76	education|school
211	19 	58	486	71	46 	583	516	europe|european|eu|e.u.
 	 	 	1	1	2	 	14	fgm|mutilation
6	1	 	11	1	14	11	34	immigrant|immigration
4	1	 	2	1	4	 	1	internet
13	148	15	1 	 	17	 	4	ireland
 	 	 	3	 	1	 	16	islam|muslim
 	 	 	1	 	1	 	 	ivory
9	12	 	6	1	6	2	2	jihad|terror
2	1	 	342	1	9	1	2 	labour
 	 	 	11	1	5	5	1	lgbt
16	1	 	33	8	39	31	5 	nhs
1	5	 	11	1	5	7	2	nuclear
 	 	 	 	 	 	 	4	nuttall
16	8	2	39	3	41	8	28	police|justice
 	 	2	3	1	2	1	3	racism|racist
 	 	 	1	 	2	1	 	saudi
3	 	 	9	1	18	296	1	scotland
 	 	2	1 	3	15	1 	8	sex
 	8	43	 	 	 	 	 	sinn
1	6	2	 	 	 	 	 	stormont
62	21	 	15	3	17	26	8	strong|stable
 	 	 	 	1	3	 	1	surveillance
 	 	 	1	 	4	6	2	syria
24	11	2	41	11	48	79	65	tax
8	 	 	3	 	12	3	1 	theresa
 	8	 	6	 	5	9	4	tourism|tourist
1	1	 	1	1	 	9	1	trident
 	 	 	1	 	2	 	2	trump
1 	1	 	7	1	18	12	3	university|universities
53	19	7	71	11	67	37	38	wealth|money|economy