Data Analysis Tutorial

cds.data_analysis loads small CSVs, normalises and smooths series, and renders ASCII plots.

Two containers, two shapes:

• DataTable — tabular data backed by string rows. load_csv returns one. Read cells with .column(name) (strings), .column_as_float(name) (floats), or .describe() for per-column statistics.
• DataSet — a list of dict rows, i.e. records rather than a grid. Its constructor takes data=, not name= / observations=.

Neither container exposes .rows or .name; use .n_rows / .n_cols and .shape / .columns.

1. Load a CSV

load_csv takes a filesystem path, so the file has to exist first. This example creates a small CSV and reads it back.

import pathlib
import tempfile

from cds.data_analysis import load_csv

path = pathlib.Path(tempfile.gettempdir()) / "cds_demo_data.csv"
path.write_text("day,temp\n1,18.0\n2,19.5\n3,22.0\n", encoding="utf-8")

table = load_csv(str(path))
print(table.headers)  # ['day', 'temp']
print(table.n_rows)  # 3
print(table.column("temp"))
print(table.column_as_float("temp"))

2. Normalise & Smooth

These take and return plain float lists. moving_average's second argument is positional (window), not a keyword.

from cds.data_analysis import normalize, z_score, moving_average

temps = [18.0, 19.5, 22.0, 21.5, 23.0]
print([round(v, 4) for v in normalize(temps)])  # min-max to [0, 1]
print([round(v, 4) for v in z_score(temps)])  # zero mean, unit variance
print([round(v, 4) for v in moving_average(temps, 2)])

3. ASCII Plots

Both plotting helpers return the rendered text, so print it to see it.

plot_bar takes a dict of label → value, not a list.

from cds.data_analysis import plot_line, plot_bar

print(plot_line(temps, title="temperature"))
print(plot_bar({"mon": 100, "tue": 110, "wed": 140}, title="sales"))

4. DataSet for record-shaped data

from cds.data_analysis import DataSet

ds = DataSet(
    [
        {"source": "SH0ES", "h0": 73.04},
        {"source": "Planck", "h0": 67.36},
    ]
)
print(ds.shape)  # (2, 2)
print(ds.columns)  # ['source', 'h0']
print(ds.column("h0"))  # [73.04, 67.36]

Run the full demo, which writes its own CSV fixture first, with:

python examples/data_analysis_demo.py