521314 / scryptool

Social Analysis of Cryptocurrencies + NLP

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

scryptool

Social analysis tool for cryptocurrencies.

Table of contents

  1. Installation
  2. Cryptocurrencies
  3. API Wrappers
    1. Santiment API
    2. LunarCrush API
      1. Metrics description
    3. Twitter API
    4. Async Twitter Scraper
    5. Other Wrappers
      1. GlassNode API
      2. Kraken-Futures API
  4. Data Analysis
    1. Market Overview
    2. Social Impact Analyzer
    3. Perspectivizer
    4. NLP for Tweets
  5. Project Organization

Installation

Create a Python 3.8 virtual environment and run the following command:

pip install -r requirements.txt

Cryptocurrencies

List of the 18 cryptocurrencies that have been analyzed.

BTC ETH BNB USDT SOL ADA XRP DOT LUNA
DOGE AVAX SHIB CRO MATIC WBTC ALGO LTC UNI

API Wrappers

1. SANTIMENT API

Method Description
list_all_coins() List all available coins in Santiment
get_price() Price (USD)
get_volume() Volume (USD)
get_daily_trading_volume() Daily trading volume (USD)
get_marketcap() Market Cap (USD)
get_social_volume() Social volume based on platform (Twitter, Reddit, Telegram, BitcoinTalk)
get_sentiment_balance() Sentiment balance of a coin
get_social_dominance() Social dominance of a coin
get_sentiment() Sentiment score (positive / negative)
get_development_activity() Development activity
get_github_activity() Github activity
get_active_addresses_24h() On-Chain Active addresses (24h)
get_exchange_balance() On-Chain Exchange_balance
get_network_growth() On-Chain Network growth
get_transaction_volume() On-Chain Transaction volume
get_perpetual_funding_rate() On-Chain Perpetual funding rate
get_circulation() On-Chain Circulation supply
get_velocity() On-Chain Velocity
get_withdrawal_transactions() On-Chain Withdrawal transactions

2. LUNARCRUSH API

Here is a short description for the LunarCrush API v2 endpoints and their respective methods. You can find more details about the request parameters here

Method Description
get_assets() Details, overall metrics, and time series metrics for one or multiple assets.
get_market_pairs() Provides the exchange information for assets, and the other assets they are being traded for.
get_market() Summary information for all supported assets (Markets page) including 5 recent time series values for some metrics.
get_global() Overall aggregated metrics for all supported assets (top of Markets page).
get_meta() Meta information for all supported assets
get_exchanges() Meta information for all exchanges that we track
get_exchange() Meta information and market pairs for a single exchange that we track
get_coin_of_the_day() The current coin of the day
get_coin_of_the_day_info() Provides the history of the coin of the day on LunarCRUSH when it was last changed, and when each coin was last coin of the day
get_feeds() Social posts, news, and shared links for one or multiple coins.
get_influencers() List of social accounts that have the most influence on different assets based on number of followers, engagements and volume of posts.
get_influencer() Individual influencer details including actual posts.

LC Metrics description

Metric Description
GALAXY SCORE The Galaxy Score™ indicates how healthy a coin is by looking at combined performance indicators across markets and social engagement. Display the real-time Galaxy Score™ of any coin.
ALT RANK AltRank™ measures a coin's performance VS. all other coins that we actively support. In general, it is a unique measurement that combines ALT coin price performance relative to Bitcoin and other social activity indicators across the entire crypto market. A coin can have a high AltRank of 1 even in a bear market situation.
INFLUENCERS View Twitter influencer activity and their impact across all coins and tokens. All influencers are measured by the same metrics, which includes followers, replies, favorites, and retweets. Metrics are evaluated across all collected posts during the timeframe selected. Actual influence will vary over time and will depend on user activity.
CANDLESTICK The incredibly powerful Candlestick widget takes any data point and compares it to price over a specified timeframe.
WORD CLOUD Uncover keywords used throughout collected social content for any coin. The Word Cloud is generated from all recent and available social posts from Twitter and Reddit. It looks at frequency of mentions. All data is segmented by either all coins or specific, individual coins.
SOCIAL FEED Display social feeds from multiple sources including Twitter, Reddit, news channels and more all at once. Gain unique insights into what's being talked about in real time. All social feeds have been cleaned with spam removed and can be organized by coin.

3. TWITTER API

Simple API wrapper for some methods from tweepy.

Method Description
get_user() Request user information.
get_all_tweets() Request all tweets given a query (Academic Research only).
get_recent_tweets() Request recent tweets given a query.
get_all_tweets_count() Request all twitter count given a query (Academic Research only).
get_recent_tweets_count() Request recent twitter counts given a query.

4. ASYNC TWITTER SCRAPER

Asynchronous Twitter scraper is based on twint project configuration. New parameters have been added to parallelize the scraping process and filter some information. The rest of the parameters can be found in the configuration file.

Additional twint parameters Type Description
Users dict Dictionary to store the coins and their respective users (parallelization purposes).
Queries dict Dictionary to store the coins and their respective queries (parallelization purposes).
Save_replies bool True to store the replies of a tweet.
Save_mentions bool True to store the mentions of a tweet.
Save_meta bool True to store all the information of the tweets.
Remove_mentions bool Remove tweet mentions.

5. OTHER WRAPPERS

Simple API Wrappers with an example endpoint request.

  • Glassnode API: Parameter parsing and requesting.
  • Kraken-Futures API: Parameter parsing and requesting.

These APIs have been considered for getting on-chain data but have not been used in the final project.


DATA ANALYSIS

You can find the complete interactive dashboard here

1. MARKET OVERVIEW

Comparison for all 18 coins based on 5 different analysis (price, market cap., social influence, on-chain and developer activity).

dashboard1


2. SOCIAL IMPACT ANALYZER

In-depth analysis of key social metrics and interactive study of correlation between price and social talk volume, per coin and month.

dashboard2


3. PERSPECTIVIZER

Tool that helps to estimate a Token A price if it had the Market Capitalization of a Token B.

dashboard3


4. NLP FOR TWEETS

Natural Language Processing of all the tweets in the last 3 months from the top 10 influencers of each coin

dashboard4


Project Organization

├── LICENSE
├── README.md
├── requirements.txt
│
├── data
│   ├── dashboard1      <- Data scraped for dashboard 1 (General Overview).
│   ├── dashboard2      <- Data scraped for dashboard 2 (Social Impact Analysis).
│   ├── dashboard3      <- Data scraped for dashboard 3 (Perspectivizer).
│   └── dashboard4      <- Data scraped for dashboard 4 (NLP for tweets).
│
├── images              <- Sample images generated from the dashboards. 
│
├── main.py             <- Main script containing all the scraping and dashboard generation pipeline. 
├── dashboard.py        <- Data processing and dataframe creation for dashboards.
├── utils.py            <- Util functions.
│
├── nlp                 <- Module for Natural Language Processing (NLP).
│   ├── parsers.py      <- Data processing for parsing and cleaning scraped tweets.
│   ├── pipeline.py     <- Pipeline for tokenization and word counting.
│   └── sentiment.py    <- Sentiment analysis functions.
│ 
├── scrapers            <- Module containing the scrapers.
│   ├── twint           <- Twint repository with custom modifications.
│   │
│   ├── __init__.py     <- Makes scrapers a Python module.
│   ├── glassnode.py    <- Glassnode simple wrapper.
│   ├── kraken.py       <- Kraken-Futures simple wrapper.
│   ├── lunarcrush.py   <- Lunarcrush wrapper.
│   ├── santiment.py    <- Santiment wrapper.
│   ├── tickers.py      <- Dictionary containing the tickers and their possible written representations.
│   └── twitter.py      <- Twitter wrappers (tweepy / parallel twint).
│
└── _config.py          <- Create your own _config.py file to store the API KEYS

About

Social Analysis of Cryptocurrencies + NLP

License:MIT License


Languages

Language:Python 100.0%