Group by Region

Example of aggregating (grouping) data on regions defined by a series with integers denoting different states.

In the figure below we are running a Steady State Detection algorithm that produces a series with two states: 0 - transient region, 1 - steady region. This binary series is used to identify the state of interest. Then we run the group_by_region calculation, specifying the state we are interested on, the type of aggregation and where the result should be placed on (timestamp).

/Users/neringaaltanaite/Project/indsl/indsl/resample/group_by.py:90: FutureWarning: The behavior of array concatenation with empty entries is deprecated. In a future version, this will no longer exclude empty items when determining the result dtype. To retain the old behavior, exclude the empty entries before the concat operation.
  result = pd.concat([result, pd.Series(data[region_index].agg(options[aggregate]), index=index)])

import os

import matplotlib.pyplot as plt
import pandas as pd

from indsl.detect import ssd_cpd
from indsl.resample import group_by_region


# Import a dataset with process data
base_path = "" if __name__ == "__main__" else os.path.dirname(__file__)
data = pd.read_csv(os.path.join(base_path, "../../datasets/data/suct_pressure_barg.csv"), index_col=0)
data = data.squeeze()
data.index = pd.to_datetime(data.index)

# Steady state detection parameters
min_distance = 60
var_threshold = 5.0
slope_threshold = -8.8

# Evaluate the Steady State Conditions
ss_map = ssd_cpd(data, min_distance, var_threshold, slope_threshold)

# Group the process data for the regions where steady state is present
aggregated_result = group_by_region(data, ss_map)

# Plot the process data
fig, ax1 = plt.subplots(figsize=(9, 7))
ax1.margins(x=0)
ax2 = ax1.twinx()
ax1.plot(data.index, data.values, label="Process data")
# Plot the aggregated result
ax1.plot(aggregated_result.index, aggregated_result.values, "ko", ms=10, label="Grouped data")
ax1.set_ylabel("Pressure (barg)")

# Plot the Steady State regions
ln2 = ax2.fill_between(ss_map.index, ss_map.values, color="orange", alpha=0.2)
ax2.margins(y=0)
ax2.set_yticks([])

# create legend below the plot
plt.legend(
    (
        plt.Line2D(data.index, data.values),
        plt.Line2D(aggregated_result.index, aggregated_result.values, color="black", marker="o", ms=10, ls=""),
        ln2,
    ),
    ("Process data", "Grouped data", "Steady State Regions"),
    loc="upper center",
    bbox_to_anchor=(0.5, -0.05),
    fancybox=True,
    ncol=3,
)
plt.title("Data grouped by region according to steady state")
plt.show()