One Wordcloud A Day

Indian state Uttar Pradesh (UP) just concluded their state election. The largest political party in India, BJP, won the election with a massive majority. The new chief minister of the state is Yogi Adityanath, who is not only a seasoned politician but also the head priest of the Gorakhnath temple in UP. There has been a lot of complaining on social media about the choice of the chief minister. I decided to analyze a few thousand tweets in order to figure out the sentiment of the masses on Twitter. The sentiment is overwhelmingly positive! There are two sentiment charts and two wordclouds that were generated one day apart. The first set was generated just after the name of the chief minister was announced and the second was generated earlier today.

The big story has been about how the cast of Hamilton gave a small speech to Michael Pence, the vice president elect of the US. More on this story is available here: I decided to use a different type of wordcloud today. In order to test it, I used a much smaller corpus of tweets, just 2000. Here is the wordcloud  Clearly people are talking a lot about Pence and the cast but Trump shows up a lot too. This is because Trump decided to attack the cast on Twitter. The sentiment around this episode is unsurprisingly negative.

The last scheduled (so far) Democratic Party debate took place couple of days back. I read news reports that said that Hillary came under fire in the debate. So I was interested in learning what happened during the debate using Twitter feed. Unfortunately, Twitter didn't discuss anything interesting! Bernie and Hillary got almost equal mentions but no issues came up as important. The only issues that showed up in the top 50 words in the wordcloud were Benghazi and immigration! It looks like people used #feelthebern hashtag often.

When I made a sentiment graph for #GE2015, which stands for General Elections 2015 in Singapore, I was surprised to see much more negative sentiment that I expected. I thought that the election would be very tight and the ruling party PAP will face major challenge from opposition parties. Yet, when the results were announced, PAP won the election with 70% vote share! The moral of the story is that in Singapore Twitter sentiment about politics is not representative of the voting population. I think that many Americans are hoping the same to be true for the USA. Super Tuesday was two days back on 1st March and the participating states included Alabama, Arkansas, Colorado (with caucuses), Georgia, Massachusetts, Minnesota (with caucuses), Oklahoma, Tennessee, Texas, Vermont, and Virginia. Additionally, Republican caucuses were held in Alaska, North Dakota, and Wyoming (https://en.wikipedia.org/wiki/Super_Tuesday#2016). Well, what do we have here? Trump is all over the wordcloud again...

Today there were two news about Sarah Palin. First, she endorsed Donald Trump for his Republican presidential bid. Second, her son Track Palin (yes, that's his name...really) was arrested for domestic violence . So I decided to use this opportunity to see which news is getting more traction on Twitter. To my surprise, Twitter doesn't care much about the son's arrest!It was Trump everywhere. However, there were a few tweets mentioning "son", "domestic", "violence", etc. The sentiment overall was highly neutral.

Today's wordcloud is for #DemDebate. It speaks for itself! Note that these are all original tweets, excluding any retweets. Looks like Bernie Sanders is getting some good coverage after all. It's not a surprise that only 3 candidate names show up all the time - Sanders, Clinton, and O'Malley. Among the issues, "healthcare" pops up frequently. The overall sentiment was neutral and positive tweets were not very high in numbers compared to the negative ones. This indicates polarized Twitter users.

Yesterday Obama gave his last State of the Union (SOTU) address and #SOTU was trending on Twitter. I downloaded 200K tweets and then analyzed only the original tweets while excluding retweets. This is now my standard format. I also reduced the number of words to show in the wordcloud. I now show only the top 50 instead of top 100 previously. The obvious words that showed up frequently are POTUS and Obama. Similarly, "president" shows up often. But there were many positive words appearing frequently - like, believe, easier, better, right, truth, strong, love, clear, good. The sentiment overall was neutral but positive tweets far outweighed negative ones.

Today's hashtag is for the Democrats debate. This was a 3-way debate between Bernie Sanders, Hillary Clinton, and Martin O'Malley in Manchester, New Hampshire. I have no idea who the third guy is. It seems that Twitter didn't care either.   The overall sentiment about the debate was mostly neutral.

Donald Trump is trending for all the wrong reasons. Today he said that Americans should not allow Muslims to enter the country! That's outright demented. But anyway, I thought it would be an interesting exercise to understand what people on Twitter are saying about him. So I downloaded close to 200K tweets and created this wordcloud. Of course, all the talk is around his comments against Muslims. But it looks to me that this is more of chatter from news outlets retweeted by people. Later you will see that the sentiment around Trump is still not as negative as you would expect. Also the derogatory words "Hitler", "dick", and "racist" are quite infrequent.   Here is sentiment graph for Trump. It's surprisingly positive. What's going on Twitter?

Sorry for the break over the last few days. I was in Shanghai and all Google services are blocked there including Blogger.com. Also, Twitter is blocked so I couldn't get any tweets to analyze. Now that I am back in Singapore, here is the wordcloud for today. Trump is at the center of the GOP debate! He is killing it (on Twitter at least). Interestingly Ted Cruz and Rubio are now showing up. Fiorina and Carson get a few mentions. But look at Jeb Bush! Can you even spot him here? The sentiment is as expected. Clearly way more positive than negative.