Bangkit 2021 Vol 3: About Bangkit Capstone Project 2021

Agung Prabowo
6 min readAug 6, 2021

--

Source : https://www.dicoding.com/blog/wp-content/uploads/2020/12/Cover.png

Hello friends. I am Agung Prabowo. Participant of the 2021 Bangkit Academy program. It has been almost a month since I graduated from the 2021 Bangkit Academy program. And this is my article that discusses my experience following the 2021 Bangkit program.

In the previous article (Vol 2), we have discussed about “My experience following the Bangkit 2021 program” Well, this time I will discuss about my personal experience in making the Capstone Project for the Bangkit 2021 program for approximately 1 month.

Table of content :

  • About Capstone Project
  • Selected Capstone Project
  • My role in the Capstone Project

Capstone Project

First we will discuss about the Capstone Project. What is the Capstone Project? In the nutshell, the Capstone Project is a final project (yes, like the final project that is done to pass D3/S1), the difference is that we are not required to make a report in the form of a paper or scientific journal. In this Capstone project task, we are required to create a product in a team of 6 people and contain elements of 3 learning paths, namely Machine Learning, Cloud Computing, and Android Developer. The products made can be in the form of mobile apps or even a tool that carries IoT (Internet of Things).

In this Capstone Project we are required to be able to provide a solution based on the 7 themes contained in Rancangan Pembangunan Jangka Menengah Nasional (RPJMN) 2020–2024, that is:

  • Economic Resilience
  • Competitive Human Resource
  • Infrastructure Development
  • National Identity and Character Building
  • Political Stability, Rule of Law, National Security & Public Services Transformation
  • Environmental Conservation, Disaster Resilience and Climate Change
  • Regional Development.

And also on the 5 priority areas in Strategi Nasional Kecerdasan Artifisial Indonesia (Stranas KA) 2020–2045. That is :

  • Healthcare
  • Bereaucratic Reform
  • Education & Research
  • Food Security
  • Mobility/ Smart City
Rancangan Pembangunan Jangka Menengah Nasional (RPJMN) 2020–2024
Startegi Nasional Kecerdasan Artifisial Indonesia (Stranas KA) 2020 – 2045

The processing time for this Capstone Project is 1 month of work. However, if it is calculated from team selection and idea design, it is approximately 2 months. Starting from the deadline for team selection on April 9, 2021, the deadline for selecting themes is April 24, 2021, then the deadline for planning is until May 3, 2021. Then, one month is given until the deadline for collection is on June 9, 2021. Actually, the deadline for collecting Caps The project was originally planned to end on June 4, 2021. However, because on May 29, 2021 after monitoring by the committee, the progress of the majority of participants to work on the Capstone Project was only about 50%. Finally, the committee extended the time until June 9, 2021.

Selected Capstone Project theme

My Capstone Project team consists of 6 people, 4 men and 2 women. The team members were myself from Singaperbangsa University, Luky Mulana from Singaperbangsa University, Dimas Kuncoro Jati from Jakarta State University, Sabrina Mutamimul Ula from Darmajaya Institute of Information and Business Lampung, Faisal Surya from Dian Nuswantoro University, and Annisa Syalsabila from Ten November Institute of Technology. This team itself is quite unique, because it consists of several different educational backgrounds such as Informatics Engineering, Information Systems, and Statistics.

Initially, our team chose between the Healthcare or Mobility/Smart City themes that we would bring for this Capstone Project. After discussions and voting between team members, we finally chose to bring the Mobility/Smart City theme for this Capstone Project. The idea of ​​our own team is especially for the tourism sector. We have an idea to make it easier for people who want to travel/travel most places easily, by showing the fastest route that can be taken to get to several places and also being able to provide recommendations for nearby tourist attractions from users or provide more personal recommendations to each user.

This is the reason why we decided to create a mobile-based application called GetLoc. GetLoc as an application that is able to recommend several tourist destinations according to the user, what the user likes, and several parameters such as city, price, category, and also time. In addition, GetLoc is also able to provide the fastest and cheapest route to visit these places, so that your travel experience will be more interesting.

GetLoc Logo
GetLoc Presentation Cover

Of course this is a promising opportunity and is also expected to provide many benefits, because in Indonesia itself there are many tourist attractions that are actually beautiful and good but are not noticed and known by many people. is expected to help advance tourism in Indonesia. If you are interested in knowing more about GetLoc, see here .

GetLoc Preview

My role in the Capstone Project

GetLoc has 3 main features. Namely Providing the fastest route, recommending places based on time, budget, and distance, then recommending based on the rating given by the user to provide a more personalized experience for each user.

GetLoc Flowchar

So my role as a Machine Learning Path is to create a Machine Learning model that can provide the fastest route to several places with the concept of TSP (Travelling Salesman Problem), TSP itself is a problem to find the minimum tour cost from a group of places. Initially our team wanted to complete this TSP using Reinforcement Learning with Q-Learning. However, after spending days but still getting results with poor accuracy, we finally decided to use Simulated annealing.

Travelling Salesman Problem with Simulated Annealing

The Machine Learning model that is then made is to recommend tourist attractions based on ratings from users, in this recommendation system we use Collaborative Filtering, which is a technique that can filter items that users might like based on reactions by similar users. It works by searching a large group of people and finding a small group of users with tastes similar to that of a particular user. We use simple DNN by embedding it first and then inserting it into 2 hidden layers, then using linear regression to predict the results.

Machine Learning model for recommendation system

Conclusion

The Capstone Project of the 2021 Bangkit Program is very extraordinary, a lot of experience and knowledge can be gained from here. This Capstone Project forces us to think fast, critically and adaptively. And also teaches us to be able to work with anyone even with people we have never met. Hopefully this article will provide a lot of knowledge, views, and also sufficient insight for the prospective participants of Bangkit 2022 later. That’s enough for now for this article about Bangkit Capstone Project, hopefully it can be useful for all of my friends who read it. Friends, you can see my article about the Bangkit program in Vol 1 and Vol 2 too.

This article has been published in Indonesian here.

Thank you.

--

--

Agung Prabowo

Interested and learn about programming, web development, and machine learning