NewsBank Text Analysis \(-\) South Africa

Section 1: China - ZAF News Articles

Keywords

1. Confucious Institute/Classroom
2. New Development Bank
3. FOCAC
4. JIMWG
5. Transnet
6. Huawei
7. Sinovac/Sinopharm
8. Eskom
9. South African Navy
10. BRICS

Frequency by Year

	2014	2015	2016	2017	2018	2019	2020	2021	2022
Chinese Media	664	860	538	988	1414	1412	763	766	265
South African Media	445	794	583	617	787	724	747	452	268
Western Media	1776	1609	1104	1660	3234	4361	3417	2622	692

20 Most Common Words

Sentiment Analysis: Time-Series Plots

Both of these plots show same information. Pick one:

Sentiment Analysis: Box Plots

Sentiment Analysis: Density Plots

Sentiment Analysis: Regression

	Dependent variable:
	sentiment
chn	0.032***
	(0.001)
wst	-0.001
	(0.001)
Constant	-0.001**
	(0.0005)
Observations	33,562
R2	0.139
Adjusted R2	0.139
Residual Std. Error	0.034 (df = 33559)
F Statistic	2,714.857*** (df = 2; 33559)
Note:	*p<0.1; **p<0.05; ***p<0.01

Interpretation:

chn = coefficient for Chinese sources
wst = coefficient for Western sources
Constant = Intercept, which is the coefficient for Domestic sources

It shows that the sentiment score for CHN-ZAF articles in domestic sources is close to zero, while Chinese sources are significantly more positive.

We can also see the same results in the boxplot and density plot above.

We can plot the above results in a coefficient plot below:

Sentiment Analysis: Coefficient Plot

Sentiment Analysis: Keywords

Both the boxplots below convey the same information. Pick one or both:

Section 2: Russia - ZAF News Articles

Keywords

1. Eskom
2. South African Navy
3. BRICS
4. Rosatom
5. Russia-Africa Summit
6. South Africa-Russia Intergovernmental Committee
7. Sputnik V

Frequency by Year

	2014	2015	2016	2017	2018	2019	2020	2021	2022
Russian Media	32	27	8	1	3	19	37	58	4
South African Media	359	572	433	505	677	621	553	340	176
Western Media	467	395	251	579	321	192	496	1223	106

20 Most Common Words

Sentiment Analysis: Time-Series Plots

Both of these plots show same information. Pick one:

Sentiment Analysis: Box Plots

Sentiment Analysis: Density Plots

Sentiment Analysis: Regression

	Dependent variable:
	sentiment
rus	-0.004*
	(0.002)
wst	-0.002***
	(0.001)
Constant	-0.003***
	(0.001)
Observations	8,455
R2	0.001
Adjusted R2	0.001
Residual Std. Error	0.033 (df = 8452)
F Statistic	5.570*** (df = 2; 8452)
Note:	*p<0.1; **p<0.05; ***p<0.01

Interpretation:

rus = coefficient for Russian sources
wst = coefficient for Western sources
Constant = Intercept, which is the coefficient for Domestic sources

We can plot the above results in a coefficient plot below:

Sentiment Analysis: Coefficient Plot

Note the difference between this coefficient plot and the coefficient plot for China - ZAF articles in the first section.

Sentiment Analysis: Keywords

Both the boxplots below convey the same information. Pick one or both:

End of File