Quick Play With EU Ref Data¶

In [37]:

#data from http://www.electoralcommission.org.uk/find-information-by-subject/elections-and-referendums/upcoming-elections-and-referendums/eu-referendum/electorate-and-count-information

In [38]:

!ls EU*

EU-referendum-result-data.csv  EURef.ipynb

Load the data in:

In [39]:

import pandas as pd
#df=pd.read_csv('EU-referendum-result-data.csv')
url='http://www.electoralcommission.org.uk/__data/assets/file/0014/212135/EU-referendum-result-data.csv'
df=pd.read_csv(url)
df[:3]

Out[39]:

	id	Region_Code	Region	Area_Code	Area	Electorate	ExpectedBallots	VerifiedBallotPapers	Pct_Turnout	Votes_Cast	...	Remain	Leave	Rejected_Ballots	Voting_for_both_answers	Writing_or_mark	Unmarked_or_void	Pct_Remain	Pct_Leave	Pct_Rejected
0	108	E12000006	East	E06000031	Peterborough	120892	87474	87469	72.35	87469	...	34176	53216	77	32	7	38	39.11	60.89	0.09
1	109	E12000006	East	E06000032	Luton	127612	84633	84636	66.31	84616	...	36708	47773	135	85	0	50	43.45	56.55	0.16
2	112	E12000006	East	E06000033	Southend-on-Sea	128856	93948	93939	72.90	93939	...	39348	54522	69	21	0	48	41.92	58.08	0.07

3 rows × 21 columns

Check the columns:

In [40]:

df.columns

Out[40]:

Index(['id', 'Region_Code', 'Region', 'Area_Code', 'Area', 'Electorate',
       'ExpectedBallots', 'VerifiedBallotPapers', 'Pct_Turnout', 'Votes_Cast',
       'Valid_Votes', 'Remain', 'Leave', 'Rejected_Ballots',
       'No_official_mark', 'Voting_for_both_answers', 'Writing_or_mark',
       'Unmarked_or_void', 'Pct_Remain', 'Pct_Leave', 'Pct_Rejected'],
      dtype='object')

How close were the votes? Count number of areas where difference between remanin and leave was less than or equal to 2 * halfwidth (i.e. both votes in band 50 +/- halfwidth.

In [51]:

halfwidth=1
df[(abs(df['Pct_Remain']-50)<halfwidth) & (abs(df['Pct_Leave']-50)<halfwidth)]

Out[51]:

	id	Region_Code	Region	Area_Code	Area	Electorate	ExpectedBallots	VerifiedBallotPapers	Pct_Turnout	Votes_Cast	...	Leave	Rejected_Ballots	No_official_mark	Voting_for_both_answers	Writing_or_mark	Unmarked_or_void	Pct_Remain	Pct_Leave	Pct_Rejected	spread
7	185	E12000006	East	E07000009	East Cambridgeshire	62435	48129	48120	77.08	48124	...	24487	38	0	18	0	20	49.08	50.92	0.08	-1.84
22	233	E12000006	East	E07000077	Uttlesford	64735	51973	51972	80.28	51972	...	26324	29	0	5	0	24	49.32	50.68	0.06	-1.36
24	252	E12000006	East	E07000096	Dacorum	108965	86313	86308	79.21	86307	...	43702	63	4	24	0	35	49.33	50.67	0.07	-1.34
25	254	E12000006	East	E07000098	Hertsmere	73295	56161	56159	76.62	56159	...	28532	34	0	12	0	22	49.16	50.84	0.06	-1.68
28	259	E12000006	East	E07000103	Watford	65060	46635	46635	71.68	46635	...	23419	49	0	23	4	22	49.73	50.27	0.11	-0.54
45	253	E12000006	East	E07000242	East Hertfordshire	106260	85446	85435	80.40	85433	...	42994	67	0	26	5	36	49.64	50.36	0.08	-0.72
49	90	E12000004	East Midlands	E06000017	Rutland	29390	22989	22986	78.20	22984	...	11613	18	0	9	2	7	49.43	50.57	0.08	-1.14
50	91	E12000004	East Midlands	E06000018	Nottingham	195394	120792	120792	61.82	120791	...	61343	130	0	47	4	79	49.16	50.84	0.11	-1.68
56	200	E12000004	East Midlands	E07000037	High Peak	72487	54864	54864	75.69	54864	...	27717	31	0	10	3	18	49.45	50.55	0.06	-1.10
61	287	E12000004	East Midlands	E07000131	Harborough	67420	54922	54918	81.44	54910	...	27850	32	0	1	2	29	49.25	50.75	0.06	-1.50
92	6	E12000007	London	E09000006	Bromley	231473	182570	182570	78.87	182570	...	90034	138	0	43	0	95	50.65	49.35	0.08	1.30
127	54	E12000001	North East	E08000021	Newcastle upon Tyne	190735	129072	129072	67.67	129072	...	63598	69	0	20	5	44	50.70	49.30	0.05	1.40
137	80	E12000002	North West	E06000050	Cheshire West and Chester	259878	193640	193633	74.51	193633	...	98082	96	0	27	11	58	49.32	50.68	0.05	-1.36
183	168	S92000003	Scotland	S12000020	Moray	71370	48148	48139	67.45	48139	...	23992	33	0	16	3	14	50.13	49.87	0.07	0.26
216	180	E12000008	South East	E07000004	Aylesbury Vale	136235	106908	106895	78.46	106895	...	53956	62	0	25	2	35	49.50	50.50	0.06	-1.00
218	182	E12000008	South East	E07000006	South Bucks	52194	40755	40755	78.08	40755	...	20647	31	0	10	5	16	49.30	50.70	0.08	-1.40
226	241	E12000008	South East	E07000085	East Hampshire	90588	73971	73971	81.65	73967	...	36576	45	0	12	1	32	50.52	49.48	0.06	1.04
248	327	E12000008	South East	E07000177	Cherwell	108342	81909	81912	75.60	81908	...	41168	72	0	29	1	42	49.69	50.31	0.09	-0.62
257	356	E12000008	South East	E07000211	Reigate and Banstead	103731	81200	81200	78.28	81200	...	40980	39	1	9	1	28	49.51	50.49	0.05	-0.98
260	359	E12000008	South East	E07000214	Surrey Heath	65569	52330	52330	79.81	52330	...	26667	25	0	8	1	16	49.02	50.98	0.05	-1.96
266	370	E12000008	South East	E07000225	Chichester	91659	71411	71406	77.91	71407	...	36326	70	3	27	1	39	49.08	50.92	0.10	-1.84
309	126	W92000004	Wales	W06000001	Isle of Anglesey	51445	37980	37981	73.82	37978	...	19333	27	0	11	0	16	49.06	50.94	0.07	-1.88
321	139	W92000004	Wales	W06000014	Vale of Glamorgan	95011	72348	72347	76.15	72348	...	35628	39	0	12	0	27	50.73	49.27	0.05	1.46
327	145	W92000004	Wales	W06000021	Monmouthshire	71607	55670	55670	77.74	55670	...	27569	40	19	0	0	21	50.44	49.56	0.07	0.88
354	58	E12000005	West Midlands	E08000025	Birmingham	707293	451422	451336	63.81	451316	...	227251	614	0	311	17	286	49.58	50.42	0.14	-0.84
368	315	E12000003	Yorkshire and The Humber	E07000165	Harrogate	119987	94669	94665	78.89	94653	...	46374	68	2	25	3	38	50.97	49.03	0.07	1.94
376	52	E12000003	Yorkshire and The Humber	E08000019	Sheffield	396406	266954	266951	67.34	266951	...	136018	198	0	76	6	116	49.01	50.99	0.07	-1.98
380	68	E12000003	Yorkshire and The Humber	E08000035	Leeds	543033	387730	387730	71.39	387677	...	192474	340	39	116	8	177	50.31	49.69	0.09	0.62

28 rows × 22 columns

Caluclate the spread(?) - difference between the remain and leave.

In [42]:

df['spread']=df['Pct_Remain']-df['Pct_Leave']

Plot the spread as a histogram, using quite tight bin widths.

In [43]:

df['spread'].plot(kind='hist',bins=200, figsize=(20,10))

Out[43]:

<matplotlib.axes._subplots.AxesSubplot at 0xac7e030c>

How about a cumulative count of the absolute spread?

In [44]:

abs(df['spread']).plot(kind='hist',bins=200, figsize=(20,10),cumulative=True)

Out[44]:

<matplotlib.axes._subplots.AxesSubplot at 0xac85d12c>

Do some nicer plots?

In [45]:

import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt

In [46]:

# Plot a historgram and kernel density estimate
sns.distplot(df['spread'], color="m",bins=100)

Out[46]:

<matplotlib.axes._subplots.AxesSubplot at 0xacd9446c>

Crib from the seaborn gallery to do a multiplot.

In [47]:

f, axes = plt.subplots(2, 1, figsize=(7, 7), sharex=True)
sns.distplot(df['spread'], kde=False, color="b", ax=axes[0],bins=200)
sns.distplot(df['spread'], color="m", ax=axes[1],bins=200)

Out[47]:

<matplotlib.axes._subplots.AxesSubplot at 0xabf7634c>

How about the absolute spread?

In [48]:

f, axes = plt.subplots(2, 1, figsize=(7, 7), sharex=True)
sns.distplot(abs(df['spread']), kde=False, color="b", ax=axes[0],bins=100)
sns.distplot(abs(df['spread']), color="m", ax=axes[1],bins=100)

Out[48]:

<matplotlib.axes._subplots.AxesSubplot at 0xabc1488c>

How about a cumulative kernel density estimate of the absolute spread?

In [49]:

sns.kdeplot(abs(df['spread']), cumulative=True)

Out[49]:

<matplotlib.axes._subplots.AxesSubplot at 0xabc1a2ac>

What if we count over the actual (signed) spread? What percentage of reporting areas voted leave (negative)?

In [50]:

sns.kdeplot(df['spread'], cumulative=True)

Out[50]:

<matplotlib.axes._subplots.AxesSubplot at 0xab95f66c>

In [ ]: