[1] 
* restate problem (comments only)
	* goal: create a household-level dataset (one row per hh_id_merge) from plot_dataset
	* input: $data/plot_dataset.dta (plot-level data)
	* output: household dataset + regression + three graphs (scatter, density, bar)

* steps (comments only)
	* 1) load plot data and keep maize plots only
	* 2) compute hh-level stats with egen (no collapse)
	* 3) tag one obs per hh and keep tagged rows
	* 4) check uniqueness + summarize + tabulate
	* 5) run regression and record results
	* 6) create graphs and export pdfs

[2] 14,177

[3] yield: mean 3227
	labor: mean 197.2
	fert: mean 103.1
	plots: mean 1.792
    irr: 292 irrigated, 13,885 rainfed

[4] sign(tot_labor) = +
    sign(tot_fert)  = +
    sign(any_irrig) = +

    Households with any irrigated plots yield, on average, 258.8 more kg than households with only rainfed plots.