Tarnished Title Lab

Inspiration Our inspiration came from a simple idea: on YouTube, a title can strongly influence whether people decide to click a video. Many creators spend a lot of time making content, but they often write titles based mainly on intuition. We wanted to build a tool that helps creators make smarter, more data-driven title decisions before publishing.

What We Learned Through this project, we learned how to combine data analysis, interface design, and AI support into one workflow. We also learned that title quality cannot be judged by only one factor, such as length. A better evaluation needs multiple perspectives.

For example, our scoring logic can be summarized as: Score=0.20S+0.25T+0.20K+0.15P+0.20M Here, S means structure and clarity, T means topic coverage, K means searchability, P means publish-time advantage, and M means similarity to historical trending videos.

How We Built It We built Tarnished Title Lab as an interactive web app. Users can choose a target market, select a video category, enter their own draft title, and set a planned publish hour. The system then compares the title with historical YouTube trending data.

The website includes a multi-factor score, AI-generated title rewrites, keyword and timing charts, and recommended historical videos ranked by relevance and popularity. We also added explanation sections such as “Why it scored,” “Rewrite direction,” and “Next moves,” so users can understand not only the score, but also how to improve it.

Challenges One challenge was making the score reasonable. At first, the system depended too much on title length, but we realized that a short title can still be effective if it is clear and specific. So we redesigned the scoring system to include topic coverage, searchability, timing, and similarity.

Another challenge was the user interface. We wanted the site to feel like a creative analysis workspace instead of a plain form. We improved the layout, charts, cards, and recommendation section to make the data easier to read and more useful.

Conclusion Tarnished Title Lab helped us understand how data and AI can support creative work. The goal is not to replace creators, but to give them evidence, suggestions, and confidence when making title decisions.