UNAIR Alumni Create Speaktograph Application

Share on facebook
Share on google
Share on twitter
Share on linkedin
FUAD and his team explain about speaktograph they created in Jawa Pos. (Photo: By courtesy)

UNAIR NEWS – Abdullah Aufa Fuad, one of the alumni of UNAIR Physics Program and Nano Science  Master’s program of University of Lyon, France has become well-known to public. He is known for his success in developing an app called Speaktograph.

In simple terms, speaktograph is as an application used to analyze speech and change the form from audio to text that can be grouped.

For Fuad, this application was designed because he felt uncomfortable seeing people often argue and claim about something. Plus, there are many experts invited on television to defend their own candidates.

“Since then, I was inspired to take part as Indonesian youth,” he said.

Fuad said, the development only took 2.5 months. He developed the speaktograph application with his friends, Kalimah Wasis Lestari (FISIP UNAIR Student) and Afifudin Maari (University of Lyon).  This speaktograph has been used to analyze two presidential debates, between two candidates Joko Widodo and Prabowo Subianto.

The results displayed by the app are in four categories. First, the talkativeness shown from total spoken words, unique words, and frequency and length of pauses.

Second, the frequent used words is about the number of repetitions of words Third, the pattern of sentences that are spoken to map out the composing words of the sentences. Fourth, characters shown from self-confidence, local wisdom, optimism, and frequency of mentioning data or numbers.

Fuad revealed, this speaktograph is different from Google Voice. The words coming out of the speaker can be calculated in one debate session. So, we will know what the speaker emphasizes more through his words.

“The results can be ranked through round graph so the data looks more interesting and easy to understand. When the size of circle is getting bigger, it means that the word is often said, “he said.

“Almost all vocabularies in KBBI have been included, familiar English has also been input. Even though there are still languages ​​that were not identified, such as a slang word selow, ” he said.

Furthermore, this application can also translate several abbreviations widely known. For example KPU, the writing appears is no longer the word Ka-Pe-U. Fuad continued that they make sure the accuracy level can reach 96 percent.

In the future, they will continue to build the program so they can track the netizens’ posts on social media. It concerns issues and products, so when it is introduced to public, they are not considered outdated.

Author: Rolista Dwi Oktavia

Editor: Khefti Al Mawalia


Berita Terkait



Media komunikasi dan informasi seputar kampus Universitas Airlangga (Unair).