Hello Dee!

I Can't seem to see a way to run a pipeline every minute, but perhaps the handle_data() function will be of use? This runs every minute. You could compile a pipeline of the Q1500, and then use the returned stocks and to calculate an RSI from these stocks. Data for this can come from the history() function.
Here is a basic algo to print every stock in the Q1500US every minute. You can add calculations from here.

from quantopian.algorithm import attach_pipeline, pipeline_output  
from quantopian.pipeline import Pipeline  
from quantopian.pipeline.data.builtin import USEquityPricing  
from quantopian.pipeline.factors import AverageDollarVolume  
from quantopian.pipeline.filters.morningstar import Q1500US  
def initialize(context):  
    # Create our dynamic stock selector.  
    attach_pipeline(make_pipeline(), 'my_pipeline')  
def make_pipeline():  
    # Base universe set to the Q500US  
    base_universe = Q1500US()

    # Factor of yesterday's close price.  
    yesterday_close = USEquityPricing.close.latest  
    pipe = Pipeline(  
        screen = base_universe,  
        columns = {  
            'close': yesterday_close,  
        }  
    )  
    return pipe  


def handle_data(context, data):  
    context.output = pipeline_output('my_pipeline')  
    for stock in context.output.index:  
        print stock  

Thank you Max!

I do see that now the pipeline output is printed every minute. But I am still concerned if we could feed data from the history() function.. From the documentation:
The most important thing to understand about DataSets is that they do not hold actual data. DataSets are simply collections of objects that tell the Pipeline API where and how to find the inputs to computations.

My understanding here is Pipelines expect references as an input, and not the actual data. On the other hand {history} function could only return the desired data.

-Dee

Hi Dee,

Your understanding of Pipeline and the fact that it expects references as input is correct. Currently, Pipeline is designed to be run once every day on daily data. Under the hood, a lot of the engineering work is optimized around this design.

To get intra-day RSI, I would recommend using pipeline to filter down the set of 8000+ equities down to a list that you want to consider trading. A good starting point is the Q1500US or the Q500US. After that, you can use data.history() to get the trailing data that you need at whatever frequency (either via a scheduled function or handle_data). You can then compute the RSI on that data.

Does this help?

Hello Jamie,

Thank you for your recommendation!

I am currently using the scheduling function every few minutes and building a pandas panel to capture the computations. Just thought an easier way would have been attaching a pipeline at fixed intervals. I will continue working in the scheduling function.

Thank you!

-Dee