Hackoclock

Evaluate content value, together with various video metrics to generate a fair value-system for TikTok to reward content creators fairly, and reimagine the ways content creators can be rewarded.

1. Running the application

1. Run 'pip install -r requirements.txt'

2. Run 'streamlit run rewards.py' for the TikTok Rewards Shop

3. Run 'streamlit run dashboard.py' for the TikTok Creator Dashboard

2. Dataset used

• Made use of dataset (https://www.kaggle.com/datasets/erikvdven/tiktok-trending-december-2020) containing fields such as creator's id, createTime, webVideoUrl, diggCount (number of likes), shareCount, commentCount etc.
• From the dataset, our own database was created with further extractions.

3. Table Documentation

creators

• creator_id: unique id of each TikTok content creator
• follower_count: No. of followers for that TikTok creator (Type: int)
• video_count: No. of total videos the creator has made
• total_points: Calculated column which sums the total points earned by the creator

videos

• creator_id: foreign key in videos table
• video_id: unique id of each TikTok video posted
• url: text string showing the link of the TikTok video (webpage)
• views, likes, shares, bookmarks, comments: Count of fields on that TikTok video
• engagement_rate: sum of likes_per_view, shares_per_view, bookmarks_per_view, comments_per_view
• likes_per_view: No. of Likes / No. of Views
• shares_per_view, bookmarks_per_view, comments_per_view: same as above but with their respective values
• date_posted: Date when the TikTok video was created and posted
• video_points: Calculated column which classifies a few variables for content creators to earn points
• sentiment: Using genai api, it helps to assign a score 0 to 1 (0 negative 1 positive) by analysing the video content through the storage buckets with the downloaded TikTok videos.

shop

Rewards for creators to redeem with the points they earned from their videos.

• reward_id: unique id for each type of reward
• amount: indicates the amount of points needed
• exposure: description of what reward is given

redemptions

Stores the log of rewards redemption by creators

• id
• reward_id
• creator_id
• amount
• exposure
• creation_date

4. Scoring Mechanism

• Using Z-score and scaling, the video_points is calculated by summing the z-scores of 7 fields: views, likes, shares, bookmarks, engagement_rate, comments & sentiment
• The points are then scaled to match the shop table requirements connected to the main webpage UI, encouraging content creators to make videos of better quality and of appropriate content.

5. Media

Creators Reward Shop

Creator Dashboard