SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World
Back to feed
Programming

Developer Builds Tool That Analyses YouTube Comment Sentiment

0
·1 views

A developer has created a personal project that fetches all comments on a YouTube video and assesses whether viewers respond positively, neutrally, or negatively. The tool classifies comments using a rule-based sentiment classifier, calculates an approval rating, and issues a verdict such as "Must Watch," "Good," "Mixed," or "Skip." It is available through three interfaces: a command-line tool, a web dashboard, and a REST API built with FastAPI. Running it requires a YouTube Data API v3 key, which the developer's public deployment does not need. The project was built partly as a way to learn backend development and handle challenges such as CORS when connecting a frontend to a backend.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingDEV Community ·

Researcher Uses Cameradar to Show Exposed IP Cameras Need Validated Access Testing

The final part of a four-part series on IP camera exposure examines how to validate whether a discovered RTSP service is actually accessible. The author used Cameradar, a tool for assessing RTSP camera infrastructure, against a camera in their own authorised test environment. An unauthenticated run found the RTSP endpoint and detected Digest authentication, but returned zero accessible streams because the credentials tested were rejected. A second run using an authorised local account and the known stream path succeeded in validating one accessible stream. The piece argues that an open port or authentication prompt does not by itself prove a compromise, and that the gap between exposure and actual access is the key point of the assessment.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Building SRT and VTT Subtitle Files from YouTube Captions in Code

A developer guide explains how to fetch YouTube captions programmatically and convert them into SRT and VTT subtitle files for editors, players, and web pages. It contrasts the two formats, noting SRT numbers cues and uses a comma for milliseconds while VTT uses a WEBVTT header and dots. The walkthrough shows a transcript API that can return both formats directly, with Python code to save paired files per video. The author notes that outputs were tested against a public YouTube transcript Actor using the first YouTube video, while the endpoint's response envelope comes from documentation. A key timing quirk is highlighted: auto-generated caption segments often overlap, so cues may need adjusting to produce valid files that players handle correctly.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Two indie hackers shut down their 3-year project Lutains, citing exhaustion and a failed pivot

Lutains, a project built over three years by two indie developers, is being shut down after thousands of euros were lost. The founders say they moved too slowly, built too many untested versions, and neglected sales. They also say their target users were too technically skilled to need a paid tool, since AI now lets them build their own. Living costs in France and the strain on their relationship and finances weighed on the decision. Despite the failure, the founders say they gained skills in offers, copywriting, branding, and AI-integrated app development.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

DataAfterlife Uses Local Gemma to Turn an Old CSV Into a Readable Story

A developer submitted DataAfterlife, a tool for the Hacktoberfest Open-Source AI Challenge that helps people find patterns and narratives in forgotten CSV files. The demo uses a fictional community cinema ledger from 1999 and 2000, with charts showing monthly lows, highs, and totals. The app runs locally, using Gemma 3 4B via Ollama to propose calculation plans and write evidence-linked stories while the browser performs the math. Users can map their own columns, switch chart types, ask plain-language questions, and export PDF reports. The project's source code and setup instructions are available on GitHub.

0 comments Read more at DEV Community