Open femtotrader opened 6 years ago
Hello,
We should have a sample data project (it could be in a separate repository)
I think it will be better to have sample data close to something that can be use in backtesting production.
I think we should split data directory from datasource and into bars, ticks, orderbooks
bars
ticks
orderbooks
I also think that we should use "round" date range for data (chunk of data)
A possible organization could be
data datasource1 bars D 2010 GOOG.csv another_ticker (to define).csv 2011 GOOG.csv another_ticker (to define).csv 2012 GOOG.csv another_ticker (to define).csv 2013 GOOG.csv another_ticker (to define).csv 2014 GOOG.csv another_ticker (to define).csv 2015 GOOG.csv another_ticker (to define).csv ticks 2015 2015-02 2015-02-01 MSFT.csv AMZN.csv GOOG.csv
This kind of organization will urge us to work with chunk of data (especially if we want to deal with tick data)
Kind regards
PS : You might have a look at https://github.com/mhallsmoore/qstrader/issues/106
Why not leverage MarketData.jl and/or Quandl.jl?
MarketData.jl
Quandl.jl
Hello,
We should have a sample data project (it could be in a separate repository)
I think it will be better to have sample data close to something that can be use in backtesting production.
I think we should split data directory from datasource and into
bars
,ticks
,orderbooks
I also think that we should use "round" date range for data (chunk of data)
A possible organization could be
This kind of organization will urge us to work with chunk of data (especially if we want to deal with tick data)
Kind regards
PS : You might have a look at https://github.com/mhallsmoore/qstrader/issues/106