I am Alfian Hakim

Scraping Project · Dashboard

Dashboard of Indonesian Government Regulations and Media Sentiment Towards Them

A dashboard aimed at examining government regulations that receive negative sentiment from the media (news). I provided the data for this dashboard by scraping the JDIH Sekretariat Negara website and Google News, and I also implemented a machine learning model to classify news sentiment.

Scraping Project · Dashboard

Dashboard of Football Academies Mapping and Hometowns of Indonesian Liga 1 Players

A dashboard that explores the mapping of football academies across Indonesia and the hometowns of Liga 1 players. I provided the data by scraping Google Maps and Transfermarkt.

Personal Project · Football Analytics

Exploring Playing Styles of Top 5 League Teams

Conducting an exploratory analysis of playing styles among football teams in the top five European leagues. I collected data from Sofascore, then identified several prevalent playing styles, and created visualizations resembling game sliders.

Personal Project · Football Analytics

English Premier League 2023-2024 Half-Season Analysis

An analysis of clubs' dominance, efficiency, and late-game drama during the 2023-2024 half-season of the English Premier League. The data was collected from FotMob.

Competition Project · Data Science

Customer Churn Prediction & Recommender System

Achieved 1st place in the Student & Junior Professional Category at Data Science Weekend 2023, organized by Data Science Indonesia & Telkomsel. The project involved exploratory data analysis (EDA), building logistic regression, linear regression, and recommendation system models based on customer cosine similarities.

Personal Project · Football Analytics

The Peak of El Pistolero: A Retro Analysis on Luis Suárez’s 2015–2016 La Liga Season

An exploratory data analysis of Luis Suárez’s best season, the season he beat two GOATs at their own playground, won the golden boot in the same league as Messi and Cristiano. The data are provided by Opta via FBRef and StatsBomb.

Personal Project · Football Analytics

Building xG and xGOT Model with StatsBomb World Cup 2022 Data

Exploring StatsBomb open data by building simple xG and xGOT model with Logistic Regression. I wrote two articles regarding this project.

Course Project · Dashboard

Sales Dashboard

Built a dashboard to monitor monthly sales of a pharmacy company. This project is done as a final task for Rakamin Academy Virtual Internship Experience (VIX) Program: Big Data Analytics at Kimia Farma.

Personal Project · Streamlit App

Letterboxd Profile Analyzer, Friends Ranker, and Movie Recommender

Letterboxd is a social media for movie lovers which they can share their movie-watching activities. This app scrapes data from Letterboxd to make visualizations and descriptive analysis based on that data. Users only need to input their username. This app has been used by more than 1800 users.

College Thesis · Data Science

Predicting MBTI Types of K-Pop Idols Based on Their Instagram Captions Using Word2Vec and LSTM

Developed deep learning models to classify MBTI types of K-Pop Idols, there's 4 models that are based on MBTI categories (I-E, N-S, F-T, J-P). Scraped the Instagram captions of 458 idols with total captions of 140.000+. The model performed better in predicting MBTI based on set of captions (each model has 0.88-0.90 macro f1-score, and 0.65 in combination of models) rather than predicting MBTI based on only one caption (each model has 0.55-0.58 macro f1-score, and 0.12 in combination of models). Deployed the models to Streamlit so that K-Pop fans can use it to predict their idols.

College Project · UI/UX Design

PUFACURE (Public Facility Quick Report)

PUFACURE is a reporting app for damages on public facility, a solution designed to improve public facility quality with the help of users. The main feature of this app is to report a damage on a public facility, with picture and post it publicly so other users can upvote it, hoping that government can immediately repair it. My role for this project is UI/UX Designer.

College Project · Data Science

Predicting Data Scientist Career Switch with Decision Tree Classifier

Implementing decision tree classifier to predict whether a data scientist switch his job or not in the future based on a Kaggle dataset. Exploring the dataset first, selecting the relevant features based on the feature correlation to the target, hyperparameter tuning the model and evaluate it with K-Fold cross validation, yielding an average accuracy of 86,9%.

College Project · Data Science

Sentiment Analysis of PeduliLindungi with Google Play Store Reviews with Naive Bayes Classifier

Building a machine learning model to classify reviews into negative or positive class using Naive Bayes algorithm. The data is gathered manually from Google Play Store, consists of 100 positive reviews and 100 negative reviews. The accuracy yielded is around 86% with a train test split ration of 75:25.

College Project · Data Science

EDA & K-Means Clustering of Disasters Affected Areas in 6 Provinces in Java

Exploring disasters happened in 6 Provinces in Java in span of 2000-2019 based on DesInventar data, looking from year to year, month to month, and which type of disaster happened most. Clustering the cities and provinces to see which area is affected most. The result shows that D. I. Yogyakarta is the most disasters affected province, related to the 2006 earthquake that suffered heavy casualties.

College Project · UX Design

Eling 2.0 (Redesigning the UX of Eling LMS)

Eling is a Moodle-based learning management system (LMS) used in Faculty of Computer Science, Brawijaya University. We were redesigning Eling with Design Thinking approach, starting from (1) Empathize by doing interviews and spreading questionnaires to understand how user's perception, (2) Define by modeling the perceptions into empathy map and defining user personas, (3) Ideate by defining ideas, requirements, information architecture, and etc., (4) Prototype by designing lo-fi and hi-fi, and building protoype, lastly (5) Test by testing the proposed prototype to users with UEQ framework.