Capstone Team CH2-PS160 [P A T A N I]

Bangkit Capstone Project 2023 Batch 2: P A T A N I (E-commerce Agriculture Application)

~ Revolutionizing Agriculture ~

· Report Bug

Table of Contents

1. About The Project
2. Documentation
   • Machine Learning
     • Library
   • Cloud Computing
     • Deployed Link
     • Repository API
     • Tools CC
   • Mobile Development
     • System Requirement
     • Installation
     • Tools MD
3. Our Team

About The Project

PATANI is a special agricultural buying and selling application that sells farm products such as vegetables, farm tools, and vegetable seeds. This application aims to facilitate accessibility between farmers and their buyers by utilizing technology. In addition, this application can be a forum that helps build an ecosystem of urban farmers. The main features are price quotes and price recommendations that are best from price predictions based on previous transactions. By using this application, it will certainly make it easier for farmers and buyers to transact, can offer prices so that they can be mutually beneficial, and are not confused in determining prices because there is already information on price recommendations for each product.

Documentation

Machine Learning

Machine Learning in the PATANI application is a price prediction for the next week where the results of the prediction are used for price recommendations for farmers and buyers. Price recommendations for farmers aim to be information that can help farmers in determining prices, while price recommendations for buyers can be used as a benchmark in submitting price offers.

For the accuracy of price predictions, we use datasets that we collect from various sources such as BPS, from Google, and prices of agricultural products in other e-commerce.

For modeling, we used Tensorflow and Long Short Term Memory (LSTM) forcasting model to predict the price of the product on the next day. To measure the accuracy, we used Root Mean Square Error (RMSE) and the average RMSE we got from our model was <0.1. For the deployment stage, we use Tensorflow Serving which is deployed to Google Cloud.

Library

• Pandas
• Numpy
• Tensorflow
• Zipfile

Cloud Computing

Created several API services to support the functionality of the application, such as login-register API, news API using a web scraper, and fish price prediction API for implementing the machine learning model. After that, we deployed them as backend services of the application on Google Cloud Platform, by utilizing Cloud Run, Cloud SQL, Cloud Storage, and Container Registry services.

Deployed Link :

• patani-backend
• patani-ml
• patani-ml-api

Repository API :

• patani-backend
• patani-ml
• patani-ml-api

Mobile Development

We carried out the process of creating static views using Jetpack Compose. Upon completion, we performed API integration with CRUD operations from the implementation of Cloud Computing APIs. With the provided API, we adjusted and implemented it within the application using Retrofit. After everything was completed, we integrated Machine Learning into the application using a model that had been exported in JSON format.

System requirement:

• Android 12+
• Android Studion Giraffe
• Internet Connection
• Java Version 1.8
• targetSDK 33

Installation:

• Download our project here Zip Application
• Build and run our project from Android Studio

Tools MD :

Our Team

• Muhammad Abdanul Ikhlas (M297BSY1103) - muhabdanulikhlas0983@gmail.com - ML
• Hisyam Agus Setiawan (M179BSY1396) - hisyamagus12@gmail.com - ML
• Zulia Amalia (M183BSX1518) - zuliaamalia108@gmail.com - ML
• Komang Yuda Saputra (C297BSY3618) - yudasaputra082@gmail.com - CC
• Bella Febriany Nawangsari (C297BSX3626) - bellafebrianynws@gmail.com - CC
• Muhamad Firdaus (A629BSY2144) - muhfrds345@gmail.com - MD
• Muhammad Nur Faiz (A123BSY2557) - mnurfaiz26@gmail.com - MD