A multi-approach analysis of the Bitcoin Blockchain

This breaks down into 2 parts: a knowledge graph and a statistical analysis.

Knowledge Graph Mapping (YARRRML to RML to RDF)

cd data

Block

yarrrml-parser -i blocks.yarrrml.yml -o blocks.rml.ttl
java -jar _rmlmapper.jar -m blocks.rml.ttl -o turtle/blocks.ttl -s turtle

Address

yarrrml-parser -i addresses.yarrrml.yml -o addresses.rml.ttl
java -jar _rmlmapper.jar -m addresses.rml.ttl -o turtle/addresses.ttl -s turtle

Transaction

yarrrml-parser -i transactions.yarrrml.yml -o transactions.rml.ttl
java -jar _rmlmapper.jar -m transactions.rml.ttl -o turtle/transactions.ttl -s turtle

UTXO

yarrrml-parser -i utxos.yarrrml.yml -o utxos.rml.ttl
java -jar _rmlmapper.jar -m utxos.rml.ttl -o turtle/utxos.ttl -s turtle

yarrrml-parser -i coinbase_utxos.yarrrml.yml -o coinbase_utxos.rml.ttl
java -jar _rmlmapper.jar -m coinbase_utxos.rml.ttl -o turtle/coinbase_utxos.ttl -s turtle

Statistical analysis

The core of this analysis can be found in the statistical_analysis notebook. This notebook gets the block data from the data_btc folder.

The btc_eth notebook contains the comparisons between the Bitcoin and Ethereum values.

Requirements

pip3 install yfinance pandas numpy matplotlib

Getting the block data

This can be automated by running the deseigned script. This script downloads the data on 5 threads concurrently to speed it up. It was notted that any more thread would result in blockchain.info DDoS protection activating and kicking all of our connections.

For the small block, at the start of the blockchain it can go as fast as 3 per second. However, for the bigger ones at the end it can slow down to 3 seconds per block.

./launch_5_downloads.sh start end

start and end are the block height at which to start downloading and the one at which to stop. It is recommented to keep both a factor of 100 to avoid rounding-related errors.

Running the notebook

Input the start and end heights selected above into the second cell of the notebook and run all.

Times of interest

We studied the range from 480000 to 540000 for the first bitcoin hype late 2017 and included this data for demonstration purposes.

We also studied and included the range from 630000 to 705000 for the effect of the chinese ban early 2020.

Saved figures

The figures used in the report are all saved in the figures folder.