panelexpr

A simple panel data evaluator

Installation

$ pip install panelexpr

Usage

Sample Data

sample_zh_2.csv:

load data

import pandas as pd
data = pd.read_csv("../data/sample_zh_2.csv")

Calculation

import panelexpr as pe
pe.eval("ma(Open, 2)", data=data, group_tag="windcode")

# equivalent to
data.groupby("windcode").rolling(2).mean()["Open"].reset_index(drop=True).rename()

Output:

0            NaN
1    1218.298745
2    1207.667866
3    1198.100075
4    1195.973899
5            NaN
6    3506.643405
7    3657.673466
8    3879.413054
9    3932.273575
dtype: float64

User defined operator

from panelexpr import TimeSeriesOperator

class MyMovingAverageOperator(TimeSeriesOperator):
    def eval(self, series: pd.Series, window) -> pd.Series: # all udo must implement this method
        s = series.rolling(window).mean()
        return s

pe.register("my_ma", MyMovingAverageOperator) # register the operator
pe.eval("my_ma(Open, 2)", data=data, group_tag="windcode")