How to get the formation energy per atom for WBM dataset?

[rydeveraumn]

Hello,

I have downloaded the WBM dataset and I am trying to figure out how to get the formation energy per atom?

Thank you in advance.

[rydeveraumn]

I was trying to reproduce the histogram under https://matbench-discovery.materialsproject.org/data. I found a readme and followed the additional instructions and clipped outlier formation energies:

[rydeveraumn]

I think I was able to figure it out.

[janosh]

glad you figured it out. for others coming here later, the data plotted in the above histogram is coming from the df_summary[MbdKey.e_form_wbm] column where MbdKey.e_form_wbm = "e_form_per_atom_wbm"

matbench-discovery/data/wbm/compile_wbm_test_set.py

Lines 453 to 479 in 93d1d5c

	e_form_hist, e_form_bins = np.histogram(
	df_summary[MbdKey.e_form_wbm], bins=300, range=(-5.5, 5.5)
	)
	x_label = {MbdKey.e_form_wbm: "WBM uncorrected formation energy (eV/atom)"}[
	MbdKey.e_form_wbm
	]
	fig = px.bar(
	x=e_form_bins[:-1], # [:-1] to drop last bin edge which is not needed
	y=e_form_hist,
	log_y=True,
	labels=dict(x=x_label, y="Number of Structures"),
	)
	fig.update_traces(width=(e_form_bins[1] - e_form_bins[0]), marker_line_width=0)
	fig.add_vline(x=e_form_cutoff, line=dict(dash="dash"))
	fig.add_vline(x=-e_form_cutoff, line=dict(dash="dash"))
	fig.layout.title = dict(
	text=f"dataset cropped to within +/- {e_form_cutoff} eV/atom", x=0.5
	)
	fig.layout.margin = dict(l=0, r=0, b=0, t=40)
	fig.update_yaxes(fixedrange=True) # disable zooming y-axis
	fig.show(
	config=dict(
	modeBarButtonsToRemove=["lasso2d", "select2d", "autoScale2d", "toImage"],
	displaylogo=False,
	)
	)