From Tweets to Thoughts – Data Analytics on 2nd Presidential Debate

March 28, 2016

Filipino viewers held on to their screens as they witnessed the second leg of Pilipinas Debate 2016.

At a time where people admittedly share almost everything, it seems that seldom would they leave their thoughts unposted in social media. Filipinos, in particular, are known as major tweeters, making trivial topics such as favorite soap operas trend worldwide.

Thus, it wasn’t surprising that the Sunday night word war of presidentiables dominated conversations on Twitter. The Commission on Elections (COMELEC) and partners themselves acknowledged this overwhelming presence of Filipinos in the social network by designating an official hashtag to centralize these tweets and also having some of them read by the hosts while waiting for the debate proper.

The second stretch of this year’s presidential debate took place in UP Cebu on March 20, from 4:30 to 9:30 PM, inclusive of a pre-debate coverage which lasted for almost two hours. The event’s official hashtag, #PiliPinasDebates2016, topped the Twitter trending list, both in the Philippines and worldwide. This time around, it garnered over 1.6 million tweets, exceeding the count of the presidential debate’s Mindanao leg of 1.3 million tweets.

Listen to the Voice

We, the Data Analytics team of Pointwest, wanted to make use of this large volume of data to draw insights regarding the event. For transparency of results, we will first provide a brief background on how we gathered the data.

#PiliPinasDebates2016 and #PiliPinasDebates were the two hashtags considered in the extraction, the latter being the one mentioned by the news anchors on TV (which could probably have misled the people from the official hashtag). Using Twitter’s streaming API and search API, we were able to collect a total of 681 233 tweets. These tweets were posted within 12:00 AM to 11:59 PM of March 20. Out of this number, 510 899 contained mentions of the presidential candidates.

We determined mentions using the terms below, as these are how the presidentiables are commonly addressed in the tweets.  Twitter has an official hashtag for every candidate, namely, #PHVoteBinay, #PHVoteDuterte, #PHVotePoe, #PHVoteRoxas, and #PHVoteSantiago which are also captured by using these keywords.

  • Binay – Jejomar, Binay, VP, Vice President, @VPJojoBinay
  • Defensor-Santiago – MDS, Miriam, Defensor, Santiago, DuRiam, @senmiriam, Meriam
  • Duterte – Du30, Duterte, @RDD_Davao, DuRiam, Digong, Rody, Rodrigo
  • Poe – Grace, Poe, @SenGracePOE
  • Roxas – Mar, Roxas, @MARoxas

Share of Voice

If the #PilipinasDebates2016 was the most talked about topic last Sunday night, which particular candidate did people talk about the most? Here we used the Share of Voice metric. A candidate’s Share of Voice is the percentage of the mentions of a particular candidate out of the total mentions of all candidates in the extracted tweets.

Twitter Share of Voice - Data Analytics

Most of the media outlets report Duterte’s popularity among social media users. However, for this debate, Binay and Roxas had the most mentions. This could have been due to the one-and-a-half-hour delay being attributed to a misunderstanding between the two camps. Duterte sat comfortably in the third place, with Poe following from a distance.

Even though Senator Miriam Defensor-Santiago was absent during the debate, her name was still one out of twenty names mentioned by tweeters.

Tone of Voice

After determining how “loud” the names of the candidates were on Twitter, the next fit question would be how the netizens talked about the candidate, whether positively or negatively.

Netizen's Reponse - Polarity - Data Analytics

We made several considerations in classifying the tweets, most importantly, on the complexity of both the Filipino language and “Taglish”, which is a common practice of mixing and matching Filipino and English words in a single sentence . Words were treated based on their context and not individually scored as positive or negative. Some of them were tagged manually as they have a different usage in the web. For example, the word “slay” has a negative connotation alone because it is associated with killing. However, when put as “Candidate slayed the debate!” it means he or she did well, hence, getting a positive score.

Even though Binay had the most mentions, most of the tweets about him were negative. Roxas received almost equal praising and bashing. Duterte and Poe, although running behind Binay and Roxas in terms of mentions, each received more positive tweets than the two of them. The number of positive tweets about Santiago merely surpassed the number of the negative ones.

Interpret the Voice

It is one thing to hear a voice, it’s another thing to understand what it says. This generation has been given an avenue to conveniently voice out their two cents, which is why it cannot be denied how important it is to hear out the unstructured chunk of views in the web by turning it into information that speaks, say in this case, the presidential candidate judged by the people as the winner of the debate.

We hope that with this little exercise by Pointwest’s Data Analytics team, we have given you a better understanding of the voices raised during the second leg of Pilipinas Debates 2016.

– – – – – – –

Founded in 2003 by pioneers of the Philippine Global Sourcing industry, Pointwest creates value for its list of satisfied clients — including top Fortune 100 and local companies — with world-class IT and BPM services backed by international-standards methodologies and innovative practises.