# Frac Fleet Data Volume CLI (Click + Pandas)

## PURPOSE
A command-line tool that quantifies the data problem Halliburton's SVP described: in hydraulic fracturing, one job per day generates about 100 GB of data streamed to the cloud, there are roughly 100 such jobs running worldwide, and that is only one of about 12 divisions. This CLI takes those fleet parameters and computes total daily and annual data volume, cloud egress/ingest cost estimates, and a "human-consumable?" verdict that shows why this scale is impossible for people to review by hand.

## WHY THIS MATTERS
The transcript makes a concrete, numeric argument for why agents are necessary: "It is impossible for human beings to consume data at this scale. And so, you need agents of some sort to do this." Before anyone builds an agent, they must first size the problem. This tool teaches that discipline — turning a leadership anecdote into reproducible arithmetic that justifies (or refutes) an automation investment.

## HOW TO BUILD
1. Use Click to define a `frac-volume` command with options for GB per job, jobs per division, and divisions.
2. Load fleet parameters from a CSV so scenarios are reproducible.
3. Use Pandas to aggregate with `.groupby("division")` and compute daily/annual totals.
4. Print a formatted table plus a per-division breakdown.
5. Add a `--format` flag for `table` or `csv` output.

Key snippet from `starter/main.py`:
```python
@click.command()
@click.option("--gb-per-job", default=100, type=int, help="GB per job per day")
@click.option("--jobs", default=100, type=int, help="concurrent jobs worldwide")
@click.option("--divisions", default=12, type=int, help="number of divisions")
@click.option("--csv", "csv_path", type=click.Path(exists=True), default=None)
def frac_volume(gb_per_job, jobs, divisions, csv_path):
    df = load_fleet(csv_path, gb_per_job, jobs, divisions)
    report = summarize(df)
    click.echo(report.to_string(index=False))
```
The summary multiplies `gb_per_job * jobs` for the global daily total and scales by annual days.

## WHEN TO USE IN PRODUCTION
Use it in data-platform capacity planning, cloud cost forecasting, storage/retention policy design, and pre-sales sizing for industrial telemetry. Do NOT use it for real-time stream monitoring or as a billing system — it is an offline estimator, not a meter.

## HOW IT HELPS STUDENTS
Students practice Click command design, option validation, Pandas groupby aggregation, and — critically — framing engineering decisions around a business-scale number. Reproducible CLI tooling is a core data-team deliverable.

## STARTER CODE
See `starter/main.py` for a Click CLI with a deterministic Pandas report and `# TODO` markers for cost rates and CSV export.
