## Monadata - Data Collection Dapp - MVP

### Motivation

Data is the most importance part of any AI application. Many AI applications struggle with insufficient data for training their models, leading to low accuracy. Monadata aims to address this issue by building a decentralized data collection platform. This platform will allow anyone to provide their data and potentially earn rewards. Additionally, users can review and verify data uploaded by others, ensuring data quality.

### User Flow

For the MVP mobile application version, Monadata will focus on an application with a lite version of the web annotation platform. The key objective is to create a simple and user-friendly interface to facilitate easy data uploading and reviewing.

**Key Features:**

1. **Login:**
- Users can log in using their EVM wallet.

2. **Data Collection:**
- **User-Generated Data:** Users can upload any picture they take in their daily life, annotate it, and post it to the system (fewer points).
- **System Tasks:** Users can submit data based on tasks provided by the system (more points).
- **Points Allocation:**
- A small number of points are awarded upon data upload.
- Additional points are granted if the data passes the review step.
- **Task Levels:** Tasks are categorized by difficulty, and users with higher points can undertake more challenging tasks (harder tasks, larger points).

3. **Data Review:**
- The review will be similar to Tinder, allowing users to swipe to mark data as correct or incorrect.
- Each data entry requires a minimum number of reviews to complete the review step.
- Data is marked as valid if more than 50% of reviewers vote it as valid, otherwise it is marked invalid.
- **Points Allocation for Review:**
- Users who upload valid data and reviewers who correctly vote receive extra points.
- Users who vote incorrectly receive negative points.

### Detailed Flows

#### Login

- **User Flow:**
1. User accesses the login page.
2. User connects their EVM wallet.
3. User signs the T&C message.

- **API Flow:**
- API to log in, verify, and store user agreements.

#### Data Collection (classification only)

**User Flow:**
- **User Upload Data:**
1. User uploads a picture and annotates it.
2. Data is posted to the system.
3. User receives initial points.

- **System Tasks:** (For system tasks, the user receives points after the task is reviewed)
1. User selects a task from the system.
2. User completes the task and uploads the data.

- **API Flow:**
1. Upload data.
- After uploading data, a review task will be automatically generated and have a deadline (e.g., 1-2 days later).
2. List tasks.
3. Upload task data.
- After uploading data, a review task will be automatically generated and have a deadline (e.g., 1-2 days later).

#### Data Review

- **User Flow:**
1. List review tasks.
2. User selects a task to access the review interface.
3. User swipes to mark data as correct or incorrect.

- **API Flow:**
1. List review tasks.
2. Get review task details.
3. Post review results.
4. Job to finish tasks:
- All tasks meeting the deadline will be finalized, points calculated, and distributed to users.
- Implement a threshold system where data is marked valid if more than 50% of votes are positive.
- Extra points for uploading valid data and correctly voting reviewers.
- Negative points for incorrect votes.

#### Technical Considerations

- **Backend:**
- Consider to use a decentralized storage solution (e.g., IPFS) for storing images.