twitter-analyzer

What it does

• Runs searches for the mentioned username(s) and then figures out
  • Most common words when people mention that given username
  • Most twitter handle pattern (e.g. CamelCase or namewithnumbers1234)

Usage

Queries

• Run a query for @twitter with cargo run --release -- query @twitter
• Run query on all accounts from conf/accounts.json with cargo run --release query

Analysis

• Analyze all stored queries with cargo run --release -- analyze
• Analyze single query for @twitter with cargo run --release -- analyze @twitter

conf/

• accounts.json: List of accounts to search (if not supplied on command line)
• analysis.json: Configuration for discarding words
  • Prepositions grabbed from https://github.com/dariusk/corpora/blob/master/data/words/prepositions.json

auth/

• Twitter API keys and tokens. Store your own bearer token under there in auth/bearer.token. Whitespace is trimmed.

Storage

• Stores queries and analyses in json with serde_json
• Storage base directory defaults to data/$QUERY
  • Analyses are stored as $SEARCH_DATE.analysis.json
  • Queries are stored as $SEARCH_DATE.query-result.json

Todo

• Remove queries from most common words
• Make an analyzer or query struct? It could have a storage handler
  • Storage dir could be configurable once instead of having to set the dir every time
• Move all logic in main to a function that returns a `Result<>