Section 1.1: Prerequisites

To begin, ensure you have the necessary libraries installed. We will utilize the pandas and Plotly graph object libraries. Plotly is a powerful open-source tool for crafting interactive, high-quality graphs.

# Install pandas

pip install pandas

# or via conda

conda install -c anaconda pandas

# Install Plotly

pip install plotly

# or via conda

conda install -c plotly plotly

Next, import the required libraries:

import pandas as pd

import plotly.graph_objects as go

Section 1.2: Data Generation

In this example, we will create our own dataset for plotting. You can substitute this with your actual data. We will create three lists:

1. X-axis data points: Years from 2001 to 2024.
2. Y-axis data points: 25 data points representing the squares of numbers from 1 to 24.
3. Standard Error (SE): For this example, we will assume the SE is one-fourth of the respective data point. Note that this is illustrative; proper statistical methods should be used to estimate SE.

def square(lst):

return [i ** 2 for i in lst]

def se(lst):

return [i / 4 for i in lst]

year = list(range(2001, 2025))

lst1 = square(list(range(1, 25)))

lst_se = se(lst1)

Section 1.3: Data Formatting for Plotting

The following function formats our lists (dates, point estimates, and standard errors). These lists are organized into a pandas DataFrame, and two extra columns for upper and lower confidence limits are computed. The function then outputs three lists:

def format_data(date, estimates, estimates_se):

cols = {'Date': date, 'Estimates': estimates, 'SE': estimates_se}

# Calculate upper and lower confidence limits

df = pd.DataFrame(cols)

df['LCL'] = df['Estimates'] - 1.96 * df['SE']

df['UCL'] = df['Estimates'] + 1.96 * df['SE']

periods = list(df['Date'])

periods.sort()

period_sorted = periods.copy()

periods.sort(reverse=True)

# Append sorted dates for plotting the confidence band

for i in range(len(periods)):

period_sorted.append(periods[i])

periods = period_sorted.copy()

LCL = df['LCL'].tolist()

UCL = df['UCL'].tolist()

UCL.reverse()

for i in UCL:

LCL.append(i)

CI = LCL.copy()

return periods, estimates, CI

Section 1.4: Creating the Plotly Graph

The Plotly library offers numerous options for graph creation. The function below constructs a Plotly graph utilizing the plotly.graph_objects.Figure class and the add_trace method. Here, we add the point estimates and the 95% confidence interval.

def plot1(y1, x1, CI_estimates, ci_dates, title, y_axis_label, x_axis_label, show_legend=True):

# Initialize figure

fig = go.Figure()

# Add Trace with point estimates

fig.add_trace(

go.Scatter(x=x1,

y=y1,

mode='lines',

hoverinfo='y+x',

opacity=0.5,

line=dict(color='blue', width=3),

name='Name of Line',

visible=True,

legendgroup="legendgroup1")

)

# Add Trace for the 95% confidence band

fig.add_trace(

go.Scatter(x=ci_dates,

y=CI_estimates,

mode='lines',

showlegend=False,

hoverinfo='y',

line=dict(color=None, width=0.10),

name='95% Confidence interval',

fill="toself",

legendgroup="legendgroup1",

fillcolor='lightblue',

opacity=0.25,

visible=True)

)

# Update layout for better aesthetics

fig.update_layout(

title_text=title,

legend_orientation="v",

font_size=14,

font={'family':'Calibri'},

title_x=0.5,

showlegend=show_legend,

xaxis_title=x_axis_label,

yaxis_title=y_axis_label,

yaxis={'linecolor': 'black', 'fixedrange': True, 'tickmode': 'auto', 'ticks': "outside", "showgrid": False},

xaxis={'linecolor': 'black', 'fixedrange': True, 'tickmode': 'array', "showgrid": False},

hovermode='x unified',

paper_bgcolor='white',

plot_bgcolor='white',

margin={'t': 125},

legend=dict(y=.00, x=0.77, traceorder='reversed'),

xaxis_range=[min(x1), max(x1)]

)

fig.update_xaxes(title_font_family="Calibri")

return fig.show()