Whom does India’s Prime Minister Follow on Twitter? A Text Mining Exercise.

Much has been said about Indian Prime Minister’s twitter following list. Quite a few number of people can be seen indulging in less than cordial discourse with people they’d identify as the opposition (which includes politicians, journalists, liberals, etc). This can be quite problematic as highlighted here —

Hall of shame — Serial abusers, sexist bigots, rumour mongers followed by PM Modi on Twitter

From following the same accounts as Narendra Modi, it appears that his personal Twitter feed is filled with rancour and distortion.

So, who’re the people being followed by India’s Prime Minister on Twitter? Is there any pattern? Although, this exercise can not conclusively (and does not intend to) answer this, it still throws up interesting results.

As of August 1st, there were around 1700 people being followed by Narendra Modi of which around 683 were Verified Accounts, while the rest (~1017) accounts were Unverified. The verified accounts include mostly official accounts, Head of Governments and other such profiles one would expect any Head of State to follow. However, the high proportion of unverified followers might raise eyebrows for some.

A simple exercise of text mining the bios of unverified people followed by Modi generated the following results —

Wordcloud For Most Commonly Occurring Words/Phrases

Some key takeaways

1. A lot of accounts (~300) seem to be quite pleased that the PM of the country is following them (as seen by count of people having “followed by” in their bios), which really isn’t surprising considering he is the most important political figure in the nation
2. Many of them identify as “hindu”, “nationalist”, “RSS”. If you had any doubts about the Prime Minister’s affiliation, it’s definitely not “Secular”, “Socialist” or “AAP”.
3. It was mildly interesting to see that most people were “blessed to be followed by PM”; a good chunk of people were “honoured” while a relatively smaller number were “proud” to be followed by our PM.
4. High count of “Gujarat” seems to suggest that PM seems to follow a disproportionate number of Gujaratis
5. Although there are a lot limitations of this particular — it’s interesting that the most common occupation (from bios of people who’ve listed theirs) is engineers (~50), followed by marketing, “professional”, writer and consultant.

Some limitations & notes

1. Many people have used Hindi in their bios and have been excluded from the exercise
2. All stop-words (commonly occurring filler words like “a”, “the”) have been excluded from the analysis
3. Many words were not counted due to typos or hashtags which would require more work
4. Many words have been clubbed together (like Modi, narendramodi, PM) under one word — Modi etc to reduce redundancy wherever possible
5. To conclude, this exercise can not be seen as representative of PM’s entire following list due to the above limitations
6. The entire study has been done using on R using the “tidytext” package.

If anyone’s interested in the code and raw data of this exercise or has any further suggestions, please let me know.