{"author_link":"\/users\/3lstmsinatrenchcoat","author_name":"3LSTMsInATrenchcoat","author_uid":"3lstmsinatrenchcoat","comments":[],"epoch":1621555446,"event":"LD48","format":"md","ldjam_node_id":257351,"likes":24,"metadata":{"p_key":"158358","p_author":"3LSTMsInATrenchcoat","p_authorkey":"1192779","p_urlkey":"378605","p_title":"Sentiment Analysis of LD48 Comments","p_cat":"LDJam ","p_event":"LD48","p_time":"1621555446","p_likes":"24","p_comments":"0","p_status":"WAYBACK","us_key":"1192779","us_name":"3LSTMsInATrenchcoat","us_username":"3lstmsinatrenchcoat","event_start":"1619222400","event_key":"109","event_name":"Ludum Dare 48"},"node":{"_collation":{"body_sanitizer":"TextUtils::SanitizeHTML via existing importer","event":"LD48","removed_author":false},"_superparent":233335,"_trust":5,"author":192779,"body":"I've always been a bit skeptical about the usefulness of sentiment analysis, but Ludum Dare comments are basically reviews, which are essentially *the* canonical setting where sentiment analysis is supposed to be meaningful, so I figured it might be interesting to take a look at the sentiment of comments from LD48.  I downloaded all comments via the ldjam.com api, and measured their sentiment using VADER via NLTK.\n\nAs you've probably already noticed, the vast majority of comments are positive:\n\n![sentiment_histogram.png](\/\/\/raw\/b01\/f2\/z\/429ec.png)\n\nThe average sentiment is 0.668 (where 1 is most positive, -1 is most negative, and 0 is neutral), and only 7.5% of comments are negative (sentiment < 0).  The sentiment is slightly lower among anonymous comments, but even then the average sentiment is 0.336 and only 21.4% of comments are negative.  Moreover, there are hardly any anonymous comments (only 112 out of 82973 comments, or about 0.13%, are anonymous).\n\nSo, given that comments are overwhelmingly positive, is sentiment (as measured by VADER) a meaningful measurement?  Well, it does seem that the average sentiment of comments on a game is correlated with the game's final score:\n\n![score_vs_comments.png](\/\/\/raw\/b01\/f2\/z\/429ed.png)\n\nSo there is a signal here, but it's not particularly strong.  However, if we restrict our analysis to games with a larger number of comments, the correlation becomes a lot higher, presumably due to the Law of Large Numbers:\n\n![sentiment_correlation_vs_num_comments.png](\/\/\/raw\/b01\/f2\/z\/429ee.png)\n\nFor comparison, here are the correlations between the overall score and various category scores:\n\n![score_correlations.png](\/\/\/raw\/b01\/f2\/z\/429f1.png)\n\nSo if you have a very large number of comments, VADER sentiment can be about as relevant to the overall score as people's actual numerical reviews in categories like Audio, Innovation, and Graphics.  But for the vast majority of games (i.e. about 90% have less than 40 ratings) the signal is about as meaningful as the Humour score, which is the least important of all category scores.","comments":14,"comments-timestamp":"2022-12-30T08:50:23Z","created":"2021-05-20T23:09:50Z","files":[],"files-timestamp":0,"id":257351,"love":24,"love-timestamp":"2021-05-24T16:42:02Z","meta":[],"modified":"2022-12-30T08:50:23Z","name":"Sentiment Analysis of LD48 Comments","node-timestamp":"2021-05-21T00:04:06Z","parent":245867,"parents":[1,5,9,233335,245867],"path":"\/events\/ludum-dare\/48\/the-company-store\/sentiment-analysis-of-ld48-comments","published":"2021-05-21T00:04:06Z","scope":"public","slug":"sentiment-analysis-of-ld48-comments","subsubtype":"","subtype":"","type":"post","version":792379},"node_metadata":{"n_key":"257351","n_urlkey":"378605","n_parent":"245867","n_path":"\/events\/ludum-dare\/48\/the-company-store\/sentiment-analysis-of-ld48-comments","n_slug":"sentiment-analysis-of-ld48-comme","n_type":"post","n_subtype":"","n_subsubtype":"","n_author":"192779","n_created":"1621552190","n_modified":"1672390223","n_version":"792379","n_status":"WAYBACK"},"source_url":"https:\/\/ldjam.com\/events\/ludum-dare\/48\/the-company-store\/sentiment-analysis-of-ld48-comments","text":"I've always been a bit skeptical about the usefulness of sentiment analysis, but Ludum Dare comments are basically reviews, which are essentially *the* canonical setting where sentiment analysis is supposed to be meaningful, so I figured it might be interesting to take a look at the sentiment of comments from LD48.  I downloaded all comments via the ldjam.com api, and measured their sentiment using VADER via NLTK.\n\nAs you've probably already noticed, the vast majority of comments are positive:\n\n![sentiment_histogram.png](\/\/\/raw\/b01\/f2\/z\/429ec.png)\n\nThe average sentiment is 0.668 (where 1 is most positive, -1 is most negative, and 0 is neutral), and only 7.5% of comments are negative (sentiment < 0).  The sentiment is slightly lower among anonymous comments, but even then the average sentiment is 0.336 and only 21.4% of comments are negative.  Moreover, there are hardly any anonymous comments (only 112 out of 82973 comments, or about 0.13%, are anonymous).\n\nSo, given that comments are overwhelmingly positive, is sentiment (as measured by VADER) a meaningful measurement?  Well, it does seem that the average sentiment of comments on a game is correlated with the game's final score:\n\n![score_vs_comments.png](\/\/\/raw\/b01\/f2\/z\/429ed.png)\n\nSo there is a signal here, but it's not particularly strong.  However, if we restrict our analysis to games with a larger number of comments, the correlation becomes a lot higher, presumably due to the Law of Large Numbers:\n\n![sentiment_correlation_vs_num_comments.png](\/\/\/raw\/b01\/f2\/z\/429ee.png)\n\nFor comparison, here are the correlations between the overall score and various category scores:\n\n![score_correlations.png](\/\/\/raw\/b01\/f2\/z\/429f1.png)\n\nSo if you have a very large number of comments, VADER sentiment can be about as relevant to the overall score as people's actual numerical reviews in categories like Audio, Innovation, and Graphics.  But for the vast majority of games (i.e. about 90% have less than 40 ratings) the signal is about as meaningful as the Humour score, which is the least important of all category scores.","title":"Sentiment Analysis of LD48 Comments","wayback_source":[]}