Explain
Plotly: enough to be dangerous
Plotly is a chart factory: you hand it a dataframe, it hands you a figure. The best plotly debugging technique is checking your dataframe before you blame the chart.
TL;DR
- Use Plotly Express (
px) first. Don’t overbuild. - Always save with
write_htmlso you can open/share. - If the plot is wrong, your dataframe is probably wrong.
The “save it to HTML” move
fig = px.bar(df, x='champion', y='games')
fig.write_html('reports/champ_pool.html')The pre-plot checklist
- Columns exist? (print
df.columns) - Sorted for time plots? (print min/max, then sort)
- No weird strings in numeric columns?
The one print that saves 30 minutes
print(df[['game_start', 'win']].head(10).to_string(index=False))