How to remove negative x-ticks on x-axis and make it start from 0?

[lolitanaz314]

Hi, it's me again. I have this code (which is a minimal reproducible example):

import pandas as pd
import numpy as np
import altair as alt

# Create data
data = {
    'channel': ['ch_0']*50,
    'media_units': [0.000000, 0.000000, 0.000011, 0.000011, 0.000022, 0.000022, 0.000033, 0.000033, 0.000044, 0.000044,
                    0.000055, 0.000055, 0.000066, 0.000066, 0.000077, 0.000077, 0.000088, 0.000088, 0.000099, 0.000099,
                    0.000110, 0.000110, 0.000121, 0.000121, 0.000132, 0.000132, 0.000142, 0.000142, 0.000153, 0.000153,
                    0.000164, 0.000164, 0.000175, 0.000175, 0.000186, 0.000186, 0.000197, 0.000197, 0.000208, 0.000208,
                    0.000219, 0.000219, 0.000230, 0.000230, 0.000241, 0.000241, 0.000252, 0.000252, 0.000263, 0.000263],
    'distribution': ['posterior', 'prior']*25,
    'ci_hi': [0.000000, 0.000000, 0.028006, 0.032422, 0.054487, 0.062808, 0.079562, 0.091343, 0.103342, 0.118192,
              0.125925, 0.143499, 0.147397, 0.167395, 0.167840, 0.189993, 0.187326, 0.211398, 0.205920, 0.231700,
              0.223682, 0.250982, 0.240667, 0.269321, 0.256925, 0.286783, 0.272501, 0.303430, 0.287437, 0.319318,
              0.301773, 0.334497, 0.315543, 0.349014, 0.328780, 0.362911, 0.341515, 0.376227, 0.353776, 0.388998,
              0.365589, 0.401256, 0.376977, 0.413032, 0.387964, 0.424354, 0.398570, 0.435247, 0.408815, 0.445736],
    'ci_lo': [0.000000, 0.000000, 0.003010, 0.003138, 0.006002, 0.006257, 0.008977, 0.009356, 0.011933, 0.012436,
              0.014872, 0.015497, 0.017793, 0.018539, 0.020698, 0.021562, 0.023585, 0.024566, 0.026455, 0.027553,
              0.029308, 0.030520, 0.032144, 0.033470, 0.034965, 0.036402, 0.037768, 0.039317, 0.040556, 0.042213,
              0.043327, 0.045093, 0.046082, 0.047955, 0.048822, 0.050799, 0.051546, 0.053627, 0.054254, 0.056439,
              0.056947, 0.059233, 0.059624, 0.062011, 0.062287, 0.064773, 0.064934, 0.067518, 0.067567, 0.070248],
    'mean': [0.000000, 0.000000, 0.009277, 0.010503, 0.018237, 0.020599, 0.026906, 0.030322, 0.035306, 0.039705,
             0.043458, 0.048774, 0.051378, 0.057552, 0.059082, 0.066060, 0.066583, 0.074317, 0.073894, 0.082338,
             0.081025, 0.090138, 0.087987, 0.097731, 0.094788, 0.105127, 0.101436, 0.112339, 0.107939, 0.119375,
             0.114304, 0.126245, 0.120536, 0.132956, 0.126642, 0.139518, 0.132628, 0.145935, 0.138497, 0.152215,
             0.144256, 0.158365, 0.149908, 0.164389, 0.155457, 0.170292, 0.160907, 0.176079, 0.166261, 0.181756],
    'channel_type': ['media']*50,
    'scaled_count_histogram': [np.nan]*50,
    'count_histogram': [np.nan]*50,
    'start_interval_histogram': [np.nan]*50,
    'end_interval_histogram': [np.nan]*50,
}

# Create DataFrame
df = pd.DataFrame(data)

# Display the DataFrame
df

def _plot_hill_curves(
    df_channel_type: pd.DataFrame,
    include_prior: bool = True,
    include_ci: bool = True,
) -> alt.Chart:

  x_axis_title = 'Media Units per Capita'
  shaded_area_title = 'Relative Distribution of Media Units per Capita'

  domain_list = [
      'posterior',
      'prior',
      shaded_area_title,
  ]
  range_list = ['#1967D2', '#D93025', '#C3C7C9']
  if not include_prior:
    df_channel_type = df_channel_type[
        df_channel_type['distribution'] == 'posterior'
    ]
    domain_list = [
        'posterior',
        shaded_area_title,
    ]
    range_list = ['#1967D2', '#C3C7C9']

  base = alt.Chart(df_channel_type).encode(
      
  )
  color_scale = alt.Scale(
      domain=domain_list,
      range=range_list,
  )
  prior_posterior_line = base.mark_line().encode(
      x=alt.X(f'media_units:Q', title=x_axis_title),
      y=alt.Y(f'mean:Q', title='Hill Saturation Level'),
      color=alt.Color(f'distribution:N', scale=color_scale),
  )
  prior_posterior_band = base.mark_area(opacity=0.3).encode(
      x=f'media_units:Q',
      y=f'ci_lo:Q',
      y2=f'ci_hi:Q',
      color=alt.Color(f'distribution:N', scale=color_scale),
  )
  histogram = base.mark_bar(color='#C3C7C9', opacity=0.4).encode(
      x=f'start_interval_histogram:Q',
      x2=f'end_interval_histogram:Q',
      y=alt.Y(f'scaled_count_histogram:Q'),
  )
  if include_ci:
    plot = alt.layer(
        histogram,
        prior_posterior_line,
        prior_posterior_band,
        data=df_channel_type,
    )
  else:
    plot = alt.layer(
        histogram,
        prior_posterior_line,
        data=df_channel_type,
    )
  return (
      plot.facet(f'channel:N', columns=3)
      .properties(
          title='Hill Saturation Curves'
      ).resolve_scale(x='independent', y='independent')
  )

_plot_hill_curves(df)

which yields the following graph:

I'm trying to get the plot to start from 0 instead of -0.00002 on the x-axis - when injecting scale=alt.Scale(domain=(0, None)) into alt.X() within prior_posterior_line to become

prior_posterior_line = base.mark_line().encode(
      x=alt.X(f'media_units:Q', title=x_axis_title, scale=alt.Scale(domain=(0, None))),
      y=alt.Y(f'mean:Q', title='Hill Saturation Level'),
      color=alt.Color(f'distribution:N', scale=color_scale),
  )

, I get a plot that looks like this which is undesirable:
.

Is there a better way to go about it such that the x-axis of the plot starts from 0? Thank you, any and all help is very much appreciated🙏

[ChristopherDavisUCI]

Hi, I didn't test this out, but does using nice=False as in this answer do what you want? #2245 (comment)

[lolitanaz314]

Hey, this worked (setting scale=alt.Scale(Nice=False) within alt.X() of prior_posterior_line)!! Thank you!!🙏🙏🙏