Data show people are tuning out of Duterte’s late-night talks

Country/area: Philippines

Organisation: Rappler

Organisation size: Small

Publication date: 13/09/2021

Credit: Dylan Salcedo


Dylan Salcedo is a data scientist at Rappler. Working in the Digital Forensics team, he designs and implements algorithms for projects that explore trends and patterns in the digital media landscape.

Project description:

Since the start of the COVID-19 pandemic, Philippine President Rodrigo Duterte has been airing weekly speeches supposedly meant to update the public on government’s pandemic response. The weekly briefings, where he is surrounded by anti-COVID task force officials, have however been used as a  platform for him to lambast his critics.  These speeches air late at night, when most people in the Philippines are fast asleep, while overseas Filipinos are expected to be awake and media are staying up for any chance of an important announcement. We analyzed if people still tune-in to the speeches as the pandemic progressed

Impact reached:

The chart we presented captured not only the viewership of the late-night speeches over time, it also highlighted events that Duterte had responded to in his speeches. Often these were rebuttals to criticisms against government pandemic response. He had attacked the emergence of community pantries that provided free food to citizens who didn’t receive aid from government, while his long-winding speeches sometimes tackled the West Philippine Sea dispute, which he blamed past administrations

Techniques/technologies used:

Livestreams of Duterte’s speeches are found in official government pages on Facebook. Metadata on posts were retrieved through Crowdtangle, a social media listening tool. Parsing and aggregating relevant data was done through Python.

What was the hardest part of this project?

The government pages had thousands of posts throughout the scanning period, and we had to develop special parsing methods to search for every livestreamed speech relevant to the story.

What can others learn from this project?

While a question can be answered with a yes or no, establishing data to support that answer not only builds the credibility of the story, it also provides further r insights into the issue.By using a data-driven approach, we were able to analyze the rare spikes in viewership, as well as spot the reasons for the downward trend

Project links: