Individual Personality Profiling Using Combination of Twitter Data and Psychological Test

IRVAN NASHER ALIMI, Mardhani Riasetiawan

2019 | Skripsi | S1 ILMU KOMPUTER

In the fields of Psychology, the main concern is to study a person’s mind which relates to their personality and profile. A post on social media can either directly or indirectly reflect someone’s personality. To gather the data required for these profiling methods, the Information Technology can contribute. The research itself will be focused on data mining and data classification with the purpose of connecting both the psychology sector and the IT sector by profiling a group of individual with data mining. The test group will consist of 50% of the total subjects that are willing to give information on their Twitter. This research goal is to create a new data set for Psychologist to analyze. Relation between data or any psychological result will not be analyzed. This research utilizes Tweepy for Twitter data collection, Briggs-Myers Test for direct Psychological evaluation process, Keirsey Temperament Sorter for classification categories, several feature extraction models (Bag-of-Words, TF-IDF, Word2Vec and Doc2Vec), WordCloud for data representation and Random Forest classifier for machine learning process. For calculation process and data showcase, Microsoft Excel is used. The result of this research are a new data set consisting of subject’s Briggs-Myers Personality Type, Twitter Classification Accuracy, Tweet Percentage and WordCloud of Tweet according to Keirsey Temperament Sorter categories.

In the fields of Psychology, the main concern is to study a person’s mind which relates to their personality and profile. A post on social media can either directly or indirectly reflect someone’s personality. To gather the data required for these profiling methods, the Information Technology can contribute. The research itself will be focused on data mining and data classification with the purpose of connecting both the psychology sector and the IT sector by profiling a group of individual with data mining. The test group will consist of 50% of the total subjects that are willing to give information on their Twitter. This research goal is to create a new data set for Psychologist to analyze. Relation between data or any psychological result will not be analyzed. This research utilizes Tweepy for Twitter data collection, Briggs-Myers Test for direct Psychological evaluation process, Keirsey Temperament Sorter for classification categories, several feature extraction models (Bag-of-Words, TF-IDF, Word2Vec and Doc2Vec), WordCloud for data representation and Random Forest classifier for machine learning process. For calculation process and data showcase, Microsoft Excel is used. The result of this research are a new data set consisting of subject’s Briggs-Myers Personality Type, Twitter Classification Accuracy, Tweet Percentage and WordCloud of Tweet according to Keirsey Temperament Sorter categories.

Kata Kunci : Twitter, Text Classification, Random Forest, Briggs-Myers Personality Test, Keirsey Temperament Sorter, WordCloud