---
title: "Comment by douzatan on \"Twitter Sentiment Analysis Using R — 2025 Edition\" | Hashnode"
description: "R still fits this kind of analysis well, especially once the visualizations matter. The sampling method should be explicit because results can shift depending o"
url: "https://hashnode.com/posts/twitter-sentiment-analysis-using-r-2025-edition/68c28edb067e2683f38cf8f1/comment/6ab32bfb35cedb6b140bb839"
---

[Discussion](https://hashnode.com/posts/twitter-sentiment-analysis-using-r-2025-edition/68c28edb067e2683f38cf8f1)

Comment [Twitter Sentiment Analysis Using R — 2025 Edition](https://hashnode.com/posts/twitter-sentiment-analysis-using-r-2025-edition/68c28edb067e2683f38cf8f1)

[douzatan](https://hashnode.com/@dzt21)

[Sep 23](https://hashnode.com/posts/twitter-sentiment-analysis-using-r-2025-edition/68c28edb067e2683f38cf8f1/comment/6ab32bfb35cedb6b140bb839)

R still fits this kind of analysis well, especially once the visualizations matter. The sampling method should be explicit because results can shift depending on whether the data came from a hashtag, a keyword, or replies to one account. Language filtering matters too. I work with AllyHub, and allyhub.com has a related social-data workflow. Did you exclude retweets, or treat them as separate observations?
