On the side of statistics and research (academic, crypto, trading), "Garbage in, garbage out" and with a same quality of data, bad variable definition and bad analytical method can cause bad results.
Some decisive factors:
- Data quality: depends on data collection or data scraping
- Variable definitions: such ones need to be presented clearly and avoid misleading intention
- Analytical methods: it is important but it is a step after data collection/ scraping and define variables
- Data interpretations: from results, from articles
Bitcoin blockchain is public so any analysis is reproducible with same data, same variable definitions, and analytic methods. The forum, websites are not academic place to publish researches so people have their reasons to hide those details but such vague articles should be read with questionable point of view.
Reproducible research. When you publish article on top journals, you have to give editors your dataset, variable definitions, dofile (or script), then they will be able to reproduce your analysis and know you are liar (fake data, fake results, ie.) or not.
Read to have info, and try to verify them, don't trust on what they present.
The article is from bitcoin[cash][dot]com.
