rschifan / emoji-www

emoji-www

Plan

Are there different emoji usages per borough / neighborhood ? Are there emojis that occur in certain areas more than others?
Are there neighborhoods which overuse or underuse emojis? If you take a topic (ie. soccer)

Objects/sentiment from image. Emoji w.r.t. Points of Interest. Create an alternative map of Manhattan based on emoji-driven neighborhoods.

Usage

• test.py script uses sample.txt

• emoji.py will try to fetch the latest emoji definitions from unicode.org and store them in a file called codepoints_latest.json

Data

pid2info, nycemoji.csv https://drive.google.com/drive/u/1/folders/0Bw7JqtQBdsZSSDdWN1VlMHVNN1E

Data format

Files chinese.tsv and english.tsv contain the extracted short canonical sequences, emojis, canonical emojis, skin tones and variations from each caption, with a Chinese and English pre-processing respectively.

Schema is: post_idcomma_separated_sequencescomma_separated_emojiscomma_separated_canonical_emojiscomma_separated_skin_tonescomma_separated_variations

Example: 977726107265881718_452803412 👭👫👬👪💑🌸🙌 👭,👫,👬,👩‍👩‍👧,👩‍❤️‍👩,🌸,🙌🏼 👭,👫,👬,👪,💑,🌸,🙌 -1,-1,-1,-1,-1,-1,2 -1,-1,-1,20,15,-1,-1 These files are meant to compute the counts.

Files chinese-tokens.tsv and english-tokens.tsv contain the Chinese and English pre-processed captions tokenized as we discussed (i.e., isolating the short canonical emoji sequences in context). Schema is: post_id<T/E>text/emoji_sequence<T/E>text/emoji_sequence<T/E>text/emoji_sequence... Example: 492186542997358343_11986392 T Pretty sure that's a smile! E 🐾🐺 T #shibatatum E 🐕🐾 These files are meant to compute the embeddings.