Workplace interaction

Import some plotting libraries and set some defaults:

import sys
import numpy as np
import pandas as pd
from IPython.display import display, Markdown
import matplotlib.pyplot as plt
import seaborn as sns
sns.set()
sns.set_style("whitegrid")

import random
random_seed = 5

sys.path.append('..')
from global_params import load_params

Tip:
Producing figures in svg format (scalable vector graphics) makes for sharp plots on webpages. However, if you make plots with thousands of observations you should set this to 'png' instead:

%config InlineBackend.figure_formats = ['svg']

Tip:
Some values apply globally to your analysis. E.g., sample sizes, cutoffs, names, rates, etc. Keeping those in a a yml file like interaction_params.yml and loading them in each notebook avoids the risk of manually adding/updating them in each notebook where they are used. You can use the load_globals function imported above from global_params.py to produce an object holding all the global values:

params = load_params('../interact_params.yml')
params
{'questions': ['Blah blah blah',
               'Blah blah blah',
               'Blah blah blah',
               'Blah blah blah'],
 'sample_size': 24}
params.sample_size
24

Sampling

subjects = pd.read_csv('../data/data_table.csv')
assert subjects.index.size == params.sample_size

Tip:
By adding a label and caption to a cell displaying a table, you can refer to that table elsewhere and insert it in a manuscript.

subjects
Table 3.1: People included in the analysis.
name age sex position nationality
0 Julie 27 F PhDstudent DK
1 Thomas 33 M Postdoc GB
2 Emilie 23 F PhDstudent CH
3 Sofie 31 F Postdoc DK
4 Sara 29 F Postdoc US
5 Cecilie 34 F Postdoc DK
6 Anders 32 M PhDstudent UK
7 Emma 42 F Professor DK
8 Caroline 31 F PhDstudent DK
9 Laura 30 F Postdoc DK
10 Mikkel 33 M Postdoc NL
11 Jens 27 M PhDstudent DK
12 Andreas 29 M PhDstudent DK
13 Jakob 28 M PhDstudent DK
14 Mathilde 61 F Professor DK
15 Katrine 35 F Postdoc DK
16 Poul 30 M Postdoc DK
17 Anna 26 F PhDstudent DK
18 Peter 42 M Professor GB
19 Ida 53 F Postdoc DK
20 Freja 30 F Postdoc DK
21 Maria 39 F Professor UK
22 Amalie 29 F PhDstudent DK
23 Camilla 35 F Postdoc DK

Tip:
By generaing markdown for descriptions that will eventually end up in the manuscript, you can imbed python values. It also ensures that the manuscript exactly reflects the notebook.

The 24 subjects from workplaces in Denmark were interviewed …. blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah

Interviews

The 24 subjects were asked to score the follow statements:

  1. Blah blah blah
  2. Blah blah blah
  3. Blah blah blah
  4. Blah blah blah

In interviewed {python} params.sample_size workplace individuals were interviewed by …. blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah,


df = pd.DataFrame({'name': subjects.name, 
                   'seniority': np.random.randint(0, 5, params.sample_size), 
                   'age': subjects.age})
df['informality'] = np.random.normal(loc=10, scale=1, size=params.sample_size)
df
name seniority age informality
0 Julie 2 27 10.061504
1 Thomas 2 33 9.795845
2 Emilie 0 23 10.704674
3 Sofie 4 31 9.995386
4 Sara 1 29 9.384324
5 Cecilie 4 34 9.617037
6 Anders 3 32 8.822115
7 Emma 3 42 8.654268
8 Caroline 0 31 10.571754
9 Laura 3 30 12.278083
10 Mikkel 2 33 10.181373
11 Jens 4 27 11.043315
12 Andreas 3 29 10.194166
13 Jakob 2 28 9.740300
14 Mathilde 2 61 8.671047
15 Katrine 1 35 10.177327
16 Poul 4 30 9.894090
17 Anna 1 26 7.958770
18 Peter 2 42 11.287880
19 Ida 0 53 10.254190
20 Freja 0 30 11.093569
21 Maria 0 39 10.223915
22 Amalie 1 29 9.613552
23 Camilla 0 35 9.975990 
sns.scatterplot(x='age', y='informality', data=df, hue='seniority', palette='viridis')
plt.ylabel('How informal you can be')
plt.xlabel('Age')
plt.legend(title='Seniority', loc='lower right', labels=['Undergrad', 'Postgrad', 'PhD', 'Postdoc', 'Professor'])
plt.ylim(bottom=0) ;
Figure 3.1: Interaction among Danes: How Danes interact is has very little to do with age and seniority, compared to most other contries.

Seems Danish people act very informally unaffected by age and seniority.

informality_age_cor = df.informality.corr(df.age)
informality_age_cor
-0.1949220780248677
informality_seniority_cor = df.informality.corr(df.seniority)
informality_seniority_cor
-0.05515869516915789

The correlation between informality and age was -0.195 and the correlation between informality and seniority was -0.055.

sns.lmplot(x='age', y='informality', data=df, hue='seniority', palette='viridis')
plt.ylabel('How informal you can be')
plt.xlabel('Age') ;
Figure 3.2: Interaction among Danes: Regressions of informality against age for five levels of seniority.