pandas integration #4

Closed
tesla1060 opened this Issue Apr 28, 2013 · 9 comments

Comments

Projects
None yet
7 participants
@tesla1060

Hi, will the lib be integrated with pandas in the future?

@gbeced

This comment has been minimized.

Show comment
Hide comment
@gbeced

gbeced Apr 28, 2013

Owner

Not in the short term.

Owner

gbeced commented Apr 28, 2013

Not in the short term.

@gbeced gbeced closed this Apr 28, 2013

@tesla1060

This comment has been minimized.

Show comment
Hide comment
@tesla1060

tesla1060 Apr 29, 2013

@gbeced i found pyalgotrade is a very flexible backtesting framework compare to the likes like zipline which is more for non-programming background users. thanks a lot. It will be great if pandas can be intergrated to further enhance the flexibility.

@gbeced i found pyalgotrade is a very flexible backtesting framework compare to the likes like zipline which is more for non-programming background users. thanks a lot. It will be great if pandas can be intergrated to further enhance the flexibility.

@gbeced

This comment has been minimized.

Show comment
Hide comment
@gbeced

gbeced Apr 29, 2013

Owner

What kind of integration are you interested in ? I mean, where would you like to have DataFrames ?

Owner

gbeced commented Apr 29, 2013

What kind of integration are you interested in ? I mean, where would you like to have DataFrames ?

@tesla1060

This comment has been minimized.

Show comment
Hide comment
@tesla1060

tesla1060 Apr 29, 2013

Hi, I was hoping that all the intermidiate data objects in pyalgotrade can be output easily, and to be timestamped.
for example

self.slow_ma = ma.SMA(feed[instrument].getCloseDataSeries(), slow_window)

I cant seem to see self.slow_ma as a timestamped data series easily. I feel it will be great convinience if every data object within pyalgotrade is kept to pandas DataFrames and Series.

Hi, I was hoping that all the intermidiate data objects in pyalgotrade can be output easily, and to be timestamped.
for example

self.slow_ma = ma.SMA(feed[instrument].getCloseDataSeries(), slow_window)

I cant seem to see self.slow_ma as a timestamped data series easily. I feel it will be great convinience if every data object within pyalgotrade is kept to pandas DataFrames and Series.

@briancappello

This comment has been minimized.

Show comment
Hide comment
@briancappello

briancappello Apr 29, 2013

Hi. I've only used Pandas a tiny bit but knowing a little about the codebase of PyAlgoTrade I can try to guess one place where it might fit. I think of DataFrames as spreadsheets on steroids (optimized in C, accessed in Python :). One of the primary use cases it's designed for is to hold time-series data; DataFrames could be used as the underlying data structure for DataSeries. This way you have one memory efficient table of data, all indexed by date (and position), with columns for OHLCV and TA indicator values. I imagine it would be a rather big undertaking and user-visible change to make the migration but, Pandas is definitely worth at least looking into!

Hi. I've only used Pandas a tiny bit but knowing a little about the codebase of PyAlgoTrade I can try to guess one place where it might fit. I think of DataFrames as spreadsheets on steroids (optimized in C, accessed in Python :). One of the primary use cases it's designed for is to hold time-series data; DataFrames could be used as the underlying data structure for DataSeries. This way you have one memory efficient table of data, all indexed by date (and position), with columns for OHLCV and TA indicator values. I imagine it would be a rather big undertaking and user-visible change to make the migration but, Pandas is definitely worth at least looking into!

@c0indev3l

This comment has been minimized.

Show comment
Hide comment
@c0indev3l

c0indev3l Sep 1, 2013

Hi,

I also would like to have Pandas lib integrated into pyalgotrade. So, we could easily feed backtester using DataFrames (or DataSeries).

Thanks for your work

Hi,

I also would like to have Pandas lib integrated into pyalgotrade. So, we could easily feed backtester using DataFrames (or DataSeries).

Thanks for your work

@ivanchee

This comment has been minimized.

Show comment
Hide comment
@ivanchee

ivanchee Jun 26, 2015

One place where a pandas dataframe could make sense is BarDataSeries. Each of the SequenceDataSeries could be be a column in a pandas dataframe. This would make the data easier to access and view/debug.

One place where a pandas dataframe could make sense is BarDataSeries. Each of the SequenceDataSeries could be be a column in a pandas dataframe. This would make the data easier to access and view/debug.

@potholiday

This comment has been minimized.

Show comment
Hide comment
@potholiday

potholiday Aug 27, 2016

Any progress in using pandas with pyalgotrade?

Any progress in using pandas with pyalgotrade?

@saitam1

This comment has been minimized.

Show comment
Hide comment
@saitam1

saitam1 Jun 26, 2017

Pandas already integrated?

saitam1 commented Jun 26, 2017

Pandas already integrated?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment