Having been a part of a similar movement during the floods, the author is feeling bad for not being able to be part of Tamil Nadu’s self help, Take II. The intent of this is to:
- analyze the uprising as it happened
- try to decipher how it happened
- who helped
- learn how can we organize ourselves better in the next ‘spontaneous’ uprising
- not take a political side, but critique poor decisions, should it arise in context.
- call out the ‘national’ media for what it is
- national media’s ‘take’ on the issue
- voice out the collective opinion of the people, which may/may not coincide with one particular POV.
- most importantly, showcase the might of people and their collective will to drive change, against political, media and geographic upheavals.
Now we begin.
Yours truly is a typical city dweller, with knowledge about Jallikattu and its traditions derived almost entirely from informal sources like Murattukalai and Virumaandi. Countryside contacts have typically been isolated to the occasional visit to best friend’s sister’s marriage close to Iyyampettai or industry visits to an investment casting foundry in Karumathampatti during engineering days. Equipped with his zero insights on Jallikattu, the author feels left out from the almost revolutionary protests in the Marina. The feeling left out bit is largely attributed to the 8000 KM distance between him and the Marina.
One good benefit of living in the digital age is that such distances do not matter to offer some help to the uprising. Especially when the help is to meddle with and dismantle data. So the author’s contribution to the protests would be to offer insights on ‘how’ it happened, what’s happening, who said what, how can we make it better next time during adversarial circumstances, and what have the ‘national’ media done, if any. All this are done with the power (read: data) of the people who’ve made it all happen.
An important aspect to note is the explicit call out of ‘adversarial circumstances’ in the above sentence. Of late, the political leadership of TN have combined indifference and inefficiency to such good effect that people had to take charge. This uprising is no different, and parts of which, will be discussed below.
Part of the challenge in getting this done is gathering the right or representative data. Like the earlier Demonetization analysis, one might think of scraping the news channels. But ‘national’ media living upto its repute, religiously covered the happenings of India and ignored the happenings of neighbouring countries like Tamil Nad! Elsewhere on Social Media, the scene was set ablaze by blitzkrieg of the knowledgeable Chennai crowd. Note: Chennai = the entirety of the Tamilsphere. ஆகுபெயர் much? So social media data. Twitter? Facebook? Facebook seem to have a greater network and reach, but scrapping data from it, especially using hashtags is a pain in certain anatomies of the body.
Twitter comes next. Ideally it’d have been good to have the entirety of the data from certain hashtags like #jallikattu, #savejallikattu, etc. But Twitter being Twitter, is coy about providing data. There is a limit on the number of and recency of tweets one can access using its API. It says ‘due to capacity constraints, the index currently only covers about a week’s worth of tweets.’ Hence, even if one wants to download tweets starting from 1st of January 2017, he/she will only get tweets from utmost a week ago. Furthermore, they say ‘Our search service is not meant to be an exhaustive archive of public tweets and not all tweets are indexed or returned.’ Therefore, there will be even less tweets available. So what’s the right or representative data? Get all the tweets we can, from 20 related hashtags. A total of 131205 tweets were gathered. There is a mandatory time gap between 2 successive tweets access using the API. So the wait was so grueling that the author was initially worried that the whole protest would be done by the time he gets hold of the data. But boy, was he proved wrong by the people!
For those who wonder what were the hashtags used for this analysis, here they are: #jallikattu, #SaveJallikattu, #amendpca, #PETA, #TamilNadu, #TNneedsJALLIKATTU, #TNneedsJALIKATTU, #WeDoJallikattu, #JusticeforJallikattu, #ISupportJallikattu, #Alanganallur, #WeNeedJallikattu, #WeWantJallikattu, #BanPeta, #AmendPCA, #Banpetaindia, #Justiceforfarmers, #Marina, #TamizhanDA, #SaveOurCultureJALLIKATTU.
It has to be kept in mind that the limit for the number of tweets is 10000 for each hashtag. Often each hashtag yielded less than 10000 tweets. As stated earlier, its not the complete set of tweets, but rather what the author was able to get out of the Twitter API (unless you;re willing to pay to get more tweets from third parties, which is course owned by Twitter!). So these might not be the most representative data, but this is what we have.
Tweets- the spontaneous uprising
Like most things great, this too had a humble beginning. A few people tweeting at the beginning, followed by a seemingly sudden increase. This is how it looks. The 2 colours represent original tweets and re-tweets. Note that this is density estimate and not count.
This temporal evolution of tweets can be combined with ‘spatial’ evolution, to find out the cities that have contributed to this dramatic increase.
One might be surprised to observe the inventive names people have for their location. This plots presents only the locations with ‘civil’ names. Chennai leads the way, followed by, India! It seems the Tamil Nadu-India neighbouring countries argument is gathering steam.
Another interesting insight is the devices people use for tweeting.
Android, iPhone and the Twitter Web Client dominate the list, but the astute observe might notice ‘SatheeshTweetBot’, arguably a product of TechSatish.
Tweets- Who made it happen and why is this important
It’d be interesting to observe who the influencers are in this network. Note that for this network visualization, only the 2 most important hashtags of #jallikattu and #savejallikattu were used to keep the network size and complexity manageable.
Even this the network of only 2 hashtags is too complex to make any meaningful observation on the hidden structures within. But there seem to be a ‘outer ring’ of end-users, beyond whom the tweets doesn’t get past. Choosing a color for each hashtag, the below plot reveal the hidden structures in the network. A keen eye might spot the ‘origins’ of each of the hashtags (though these origins could well be the origins in the time frame of the tweets gathered, probably not the whole tweet network)
Removing the ‘dead-end’ users with whom a tweet doesn’t propagate in their network further and selecting the users whose tweets reach 5 or more users, a more manageable network appears.
Not only does this reveal such patterns within, it also also helps identify the most influential people in the network of 2 related hashtags. This identification of influencers is critical to the success of the future uprisings, as they have greater reach and influence in this network and spreading the word around would be most efficient originating from these influencers. Nodes 925 and 865 seem to be the most influential users of their hashtags respectively. Your’s truly has intentionally withheld their names for the sake of privacy.
Another reason its vital to have these influencers benchmarked is to help in times of adversity. Remember during the floods, nothing remotely resembling a working government was around. But somehow help reached people. Who made it happen? How did the word get around? And even during this uprising, in the initial phase when the govt was pretty much non-existent and media put u-turn on us, word did get around. And often these influencers are not popular figures. It could be a common man! Getting to know who they are and their reach would help mobilize people when the ‘usual’ way of getting info from governments and media fail. One might argue that this is not organic growth. True. But its as close to being organic without being organic, in situations that should be (in the future) called ‘what the paneer‘.
While the network describes the flow of information across, given the emotional connect of the protests, it’d be just as interesting to decipher the emotions of the people/tweets. For someone like the author, who is far from ground zero, it gives the best insights/feels of the souls protesting back home. A grim ban on tradition. How people respond?
For a second the author took jingoistic pride on being a Tamil, before returning to ground state. It’s morale-boosting, perumai-coming, kaneer-vandhing story.
National media against people’s power
So they did what they are made/paid for. Cover the nation. Not neighbouring places Tamil Nad.
All of them, identically and unanimously, dropping that alphabet reveal that they have no interest in anything news related to words ending with that alphabet. Examples being tamilnadU, yoU, jallikattU. By national media, the author refers to those same agencies that were featured in the demonetization discussion, NDTV, CNN IBN, Times Now. So what was their coverage of Jallikattu?
Just around Pongal, they seem to have ‘some’ articles related to Jallikattu. So when the uprising began, there has actually been a dip in coverage. Tamil folks are not new to this. So what do they do?
They make their own news! Its the same graph as before, but with the number of mentions of jallikattu on twitter added. If at all one needs a good example of ‘off-the-charts’, this is it. People with a combination of tremendous patience and eyesight might spot the purple Times Now line going up a touch on the 18th of Jan. But you know who the real star of this plot is. A true TamizhanDa moment.
While as a reader you might wonder what topics these channels covered. So here they are:
So CNN IBN’s main topic of discussion seem to be PETA, especially during the protest days!
Sentiment Analysis reveal that Times Now has a different spin to its articles.
And of the 3 channels, only Times Now seem to feature a high percentage of overall negative emotion in their news content. Wonder if they’re trying to take a so-called moral high ground? Of late Times Now has drawn flak from the Tamil community for insensitive comments on the issue, and analysis of their news data suggests why they incurred the wrath of the Tamil community.
The author started the day on a disappointing note, because of being unable to join the protests in front of the Indian Embassy in The Hague. But the mood improved to inspiring after finishing the analysis that spanned almost a week. The highlights were:
- The uprising started spontaneously, evolved organically that resulted in 1 in every 8 homo sapiens in the city gather at Marina. Yes, half a million citizens gathered at the marina.
- Having the second longest beach in the world has always been a pride, but never been put to a better use than today.
- Many many cities and communities have voiced their support against ban of an age old tradition, perpetrated by an organization with questionable standards and ethics.
- We have a beautiful and committed network of commen men/women/children who made this happen.
- We have a world-wide network.
- Getting large quantities of data from twitter is a real pain.
- In future what-the-paneer situations, we now know whom to go to, to have the maximum and fastest reach in the protesting network.
- Peace loving, knowledgeable Chennai crowd has lived upto its repute. Sentiment analysis reveal they also own the tag of being-joyous-and-positive-even-during-protests.
- National media is good in being national media, but terrible when it comes to covering southern issues.
- CNN IBN seem to like PETA.
- Times Now doesn’t seem to like Jallikattu.
- In a truly kannula-thanni moment- we make, cover and report our own news.