Today,17th September 2017, is Prime Minister Modi’s 67th Birthday. While there will be many discussions on aspects of his personality and his policies, we at ‘The True Picture’ decided to take a different approach and explore a topic which lends itself to frequent debates – what is the reason behind the enduring popularity of Modi? Remember, this year Modi would be completing 16 years in a row as the head of a form of democratically elected government – from 7th October 2001 till May 2014 as the Chief Minister of a large state and from 26th May 2014 as Prime Minister. By any standards, nationally and internationally, this is a significantly long stretch of sustained popularity for a politician in a democratic society. As recent surveys, my many media houses, indicate Modi’s popularity is not only sustaining but indeed increasing even beyond the 2014 levels when he won the national mandate.
While normal journalistic practice is to answer such a question through the prism of current political context in the country, and indeed such an exercise is indulged in by many reputed columnists every year, we have approached the problem differently – through the tools of data science.
The government led by Narendra Modi completed three years in office on 26th May 2017. During those three years, Modi has had the opportunity to deliver umpteen number of speeches – in official functions and obviously in electoral rallies. Is there a pattern, a design, a focus, an insight that we can draw from all of these speeches? These three years have seen the Prime Minister launch several key initiatives, across sectors, such as the Pradhan Mantri Fasal Bima Yojana, Pradhan Mantri Mudra Yojana, Start-up India, Make in India, Digital India, Jan Dhan Yojana, Ujjwala Yojana, Beti Bachao Beti Padhao, Sukanya Samriddhi Yojana etc. Swachh Bharat is an important national movement that was also started by the present government.
In addition to launching various flagship schemes, there are various other occasions when the intent and the direction that the government is taking is communicated through speeches. And who better to under the direction of the government than the head of government?
This was an intriguing question that excited our data science team when the question was first posed to them. After an animated debate, we settled for the following methodology.
First, we would analyse, using Natural Language Processing (NLP) techniques, only the official speeches of Modi during his first three years in office. Electoral speeches are by occasion more political and therefore mixing them with official speeches will skew the data (although this itself can be a subject of a separate study). Second, the period of analysis was chosen from June 2014 to April 2017 -the first three years of the government. This would set a base for a sufficiently large number of speeches to be analyzed, thereby giving an opportunity to draw insights on the first three years as well as analyzing patterns that may continue in the fourth year. In all, 535 speeches delivered by the Narendra Modi during this period were thus chosen for analysis.
Such an objective analysis of the speeches involves determining the frequency of the words and to use modern methods to estimate the structure of the speeches. This method is particularly useful for analyzing a large amount of text or a great number of speeches together. This sort of analysis is also useful because it is relatively unaffected by personal opinions and goes beyond the usual rhetoric to quantify the words being used in the speeches.
Method 01: Machine Algorithms Results
A) Frequency of Words
The choice and use of terms spoken by those in the public eye are generally not by chance but by choice and design. The study of the absolute frequency of words provides a clear indication on the message being conveyed.
For this study, some words were filtered out for natural language processing of the text (also known as stop-words) which include common words such as “is”, “are” etc. and words which are filler words such as “brothers and sisters” or “my fellow countrymen”, etc.
After the stop-words were filtered out, we looked at the frequency of words in the totality of speeches of three years, to discern what has been the focus of the speeches.
In the charts below, X axis represents the words; Y axis represents the absolute count.
The top word count was also done year-wise to assess if there has been any change in focus between the years? Below are the results for each year.
The results are both remarkable and startling in what they reveal. In the course of over 535 speeches in totality over three years, and when looked at in batches year-by-year, the results are unwaveringly the same – development, poor and power(electricity) have been the overarching theme in the discourse month after month, year after year. If we look at the top 5 words, they in totality present a narrative that the consistent focus has been only on development and economic issues with emphasis on improving the ‘quality of life of the poor’
B) Word Cloud – Speeches of all three years
Word cloud is generated using machine algorithm, which places words in context and discerns emphasis on the words in the way sentences are framed around that context. Words clouds thus offer an additional insight on emphasis apart from pure frequency count. The way to read the word cloud finally generated is this – the larger the size of a particular word, the higher the emphasis on that word in the corpus of speeches analyzed. The word cloud of all the 535 speeches is below.
‘Transform’ and ‘Future’ are two words that are quite prominent in the overall word cloud, in addition to other words that are prominent such as ‘Crop’, ‘Mudra’, ‘Farmer’, “gigawatt’, etc. What this tells us is that in so much as the speeches are concerned, a great amount of emphasis has been placed on talking to the audiences about the transformative agenda for the future.
Method 02: Classification in Categories
To assess the key areas which were emphasized in the speeches, a corpus or a dictionary of sorts was created, after removing stop-words (like ‘the, ‘a,’, ‘an’, etc.), containing common governance terms and then they were categorized into representative categories. For example, all words related to agriculture such as farmers, fertilizers, soil health, crop insurance, etc. were categorized into Farmer Welfare. After creating this corpus, a monthly frequency of words was calculated for what emerged as focus areas of the speeches. Below are some of the trends that emerged from these speeches.
A) Farmer Welfare and Rural Development
Agriculture plays an important role in the Indian Economy and therefore the welfare of farmers and the development of rural areas is something all governments endeavor to work on. The graph for this section depicts the emphasis on these areas in the text of the speeches.
The graph indicates that there has been consistent and sustained focus on agriculture, farmer and rural development issues. The graph peaks in February 2016. Narendra Modi launched the National Rurban Mission in Chhattisgarh as well addressed the Kisan Kalyan Mela this month.
‘We are committed to provide proper healthcare facilities and
good education to children in our villages’
Modi at the launch of the National Rurban Mission
‘Welfare of the farmers is at the core of Pradhan Mantri Fasal
Modi at the Kisan Kalyan Mela
On an average, we can see that the words related to ‘Farmer Welfare and Rural Development’ were mentioned 143 times monthly, the highest being in February 2016 and the lowest in December 2014.
B) Infrastructure Development
Developing the infrastructure of the country is nominally a priority of all governments. But what about the current government? As we can see, infrastructure related words have been mentioned on an average 205 times monthly over the last years.
‘Power, water, roads pre-requisite for development’
Modi at the launch of Hydro Electric Power Project and Highway
Project at J&K in November 2015.
Looking at the number of times that words related to Economy got mentioned, we see that it was mentioned on an average 147 times monthly. The graph below depicts the monthly trend of the economy related words over the last three years.
We can see a sharp spike in November 2015. This was an important month from the country’s perspective as we saw the Modi travel to Turkey for the G20 Summit, as well as for the BRICS leaders’ meeting held on the sidelines of the G20 summit. During this month, Modi also travelled to Britain where he addressed the British Parliament. November 2015 also saw the ASEAN-India Summit in Kuala Lumpur and Modi’s address to India-Singapore Economic Convention. During this month, Modi also spoke at many occasions domestically such as the Delhi Economics Conclave or the at the Sher–E–Kashmir Stadium, Srinagar.
‘G20 must align itself with the SDGs to stimulate faster and a
more broad-based economic growth’
PM at G20 Working Lunch on Development and Climate change,
Here we can see that July 2015 saw a spike. This is the month during which Digital India was launched. Then in September 2015 we saw a special emphasis on Technology related words. On examining the speeches, we can see that this month the Modi spoke at the Digital India and Digital Technology Dinner held at San Jose, California as well as the Townhall Q&A at Facebook HQ, San Jose. He also spoke to the Indian Diaspora at the SAP centre San Jose. This month also saw Modi talk at the UN Peacekeeping Summit, UN Summit for adoption of Post-2015 Development Agenda as well as his opening remarks at the G4 Summit in New York.
In India, the Prime Minister addressed a special session of National Meet on Promoting Space Technology that month, thus September 2015 sees a notable spike in technology related words.
‘Facebook, Twitter, Instagram are the new neighborhoods of our
Modi at the Digital India and Digital Technology Dinner, San Jose, California.
E) Skill and Education
In all references to Skill or education were made on an average 165 times monthly over the past three years.
‘Matching job creation with industry demand is the key to
Modi at the launch of the Skill India Mission in July 2015.
In a vibrant and functional democracy, the political leaders are assessed by the citizens through a multitude of factors. Credibility is one such matrix and something on which most serious politicians want themselves to be ranked fairly high in the measuring scale. Credibility in as much as the topics and domains they focus their energies on as it is through the consistency of their words and actions. As the results above establish, what is remarkable about Modi is his persistent and consistent focus on the topics of development and poor. Month after month, year after year, in normal times and in times of stress and campaigns, the data suggests that his focus from development issues has not wavered. As political scientists’ asses the phenomenon of Modi, this is one dynamic they can perhaps factor in while analyzing why he arguably remains so popular despite now being three years in power at the centre (and before that for 12 years year in Gujarat) and why anti-incumbency never seems to afflict his political prospects. Is it the ‘credibility factor’ about Modi – the unwavering focus on only and only developmental issues even as others around him hopping from one topic to the other – that keeps his popularity so high? And therein perhaps also lies an answer as what in the coming months and years!
The speeches were collected from the website http://www.narendramodi.in/ . Many of the speeches were originally in Hindi. They were first translated in English and then the machine algorithms were applied for study.