Mevel - Bangkit Capstone Project Presentation

Summary

Mevo is an innovative mobile app designed to promote ecotourism in Indonesia by using machine learning to recommend hidden tourist destinations based on user-uploaded images. Despite Indonesia’s vast natural resources and 583 islands, many local and international tourists remain unaware of these attractions. Mevo leverages TensorFlow, Firebase, and Android Studio to provide personalized travel suggestions, with a focus on ecotourism. The app offers a freemium model targeting both B2B and B2C markets, aiming to boost local economies, promote sustainability, and reduce global warming. It plans to expand through pilot projects and partnerships over the next six months.

Takeaways

• 😀 Indonesia has numerous natural attractions, but only 21 of them are widely known both locally and internationally.
• 😀 The problem is that 67% of the Indonesian population is unaware of the country's tourism potential, leading to underutilization of resources.
• 😀 The proposed solution is Mevo, an innovative tourism app that uses machine learning to recommend destinations based on user-uploaded photos.
• 😀 Mevo utilizes TensorFlow for its machine learning model, using MobilenetV2 to classify and recommend destinations based on images.
• 😀 The app is built using Android Studio, with Firebase managing backend services such as cloud functions, data storage, and authentication.
• 😀 Mevo promotes ecotourism by recommending lesser-known, eco-friendly tourism spots and supports local communities through tourism revenue.
• 😀 The app's machine learning model has achieved a 70% accuracy rate, making it a promising MVP (Minimum Viable Product) for future development.
• 😀 Mevo outperforms competitors like Foursquare and Expedia by offering image-based recommendations and promoting ecotourism.
• 😀 The app’s market strategy targets both B2B (business-to-business) and B2C (business-to-consumer) markets, focusing on demographic and psychographic segmentation.
• 😀 Mevo plans to expand through social media, partnerships, and pilot projects, with a goal of educating the public and raising awareness about tourism opportunities.
• 😀 In the next six months, Mevo aims to continue its development with a core team of six members, with a focus on enhancing the app's features and user experience.

Q & A

What is the main goal of the Mevo application?

The main goal of Mevo is to promote lesser-known tourist destinations in Indonesia by providing recommendations based on images uploaded by users. It aims to increase awareness of these hidden gems, ultimately boosting local tourism and economic growth.

How does the Mevo app generate destination recommendations for users?

Mevo uses machine learning, specifically a model built with TensorFlow and MobileNetV2, to analyze images uploaded by users. The app then provides destination recommendations based on the content of these images, helping users discover relevant tourist spots.

What machine learning technologies are used in Mevo?

Mevo utilizes TensorFlow for machine learning and employs the MobileNetV2 model, which is a pretrained deep learning model. The model is fine-tuned with a dataset to identify and recommend destinations based on the images provided by users.

What is the accuracy of the machine learning model used in Mevo?

The machine learning model used in Mevo achieved an accuracy of 70%, which is considered a strong result for an initial minimum viable product (MVP). The accuracy measures how effectively the model can classify and recommend destinations based on the uploaded images.

What technologies are used for the app’s backend and database management?

The backend of Mevo is powered by Firebase, a cloud-based service provided by Google, which handles authentication, database storage, and cloud functions. Firebase Cloud Functions are used to manage the app's server-side logic, allowing for scalability and ease of maintenance.

How does Mevo differentiate itself from other tourism applications?

Mevo differentiates itself by using an image-based recommendation system powered by machine learning, unlike traditional tourism apps that typically rely on search categories. It also focuses on promoting ecotourism and aims to increase local economic development while supporting environmental sustainability.

What is the app’s approach to promoting ecotourism?

Mevo promotes ecotourism by recommending destinations that follow sustainable tourism practices. It aims to help local communities develop ecotourism areas that support environmental preservation while also providing economic benefits, such as job creation and local business growth.

What are the two main market segments that Mevo targets?

Mevo targets both B2B (business-to-business) and B2C (business-to-consumer) markets. For B2B, it focuses on partnerships with tourism agencies and other institutions, while for B2C, it aims to reach individual users through social media and advertisements.

What business model does Mevo follow?

Mevo follows a freemium business model, where basic features are free for users, and premium features are available through a subscription. It also generates revenue through advertising and potential partnerships with other tourism-related businesses.

What are the future plans for Mevo over the next six months?

Over the next six months, Mevo plans to expand its features, improve its machine learning models, and collaborate with local tourism agencies for pilot projects. The goal is to increase the app's reach, improve user engagement, and further promote ecotourism in Indonesia.