Plotly: Plotting a time series using plotly and datetime index

Question

I am trying to plot a line graph of the yield spread from 1993 to 2020 using plotly. The name of the variable is "yieldsp" in the dataframe called "data". It is a time-series data with a DateTime index as follows:

data['yieldsp'].head()

Date
1993-10-01    2.36
1993-10-04    2.32
1993-10-05    2.29
1993-10-06    2.31
1993-10-07    2.28
Name: yieldsp, dtype: float64

data.index

DatetimeIndex(['1993-10-01', '1993-10-04', '1993-10-05', '1993-10-06',
               '1993-10-07', '1993-10-08', '1993-10-12', '1993-10-13',
               '1993-10-14', '1993-10-15',
               ...
               '2020-06-12', '2020-06-15', '2020-06-16', '2020-06-17',
               '2020-06-18', '2020-06-19', '2020-06-22', '2020-06-23',
               '2020-06-24', '2020-06-25'],
              dtype='datetime64[ns]', name='Date', length=6688, freq=None)

I wrote the following code to get the plot:

# Using plotly.express
import plotly.express as px

#data = px.data.iris()

fig = px.line(data['yieldsp'], x = data.index, y ='Yield Spread', line_shape="spline", render_mode="svg")
fig.show()

But it generated the following error:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-49-92ef77a6fd5a> in <module>
      5 
      6 fig = px.line(data['yieldsp'], x = data.index, y ='Yield Spread', color="continent", line_shape="spline", 
----> 7               render_mode="svg")
      8 fig.show()

~/opt/anaconda3/envs/tsa_course/lib/python3.7/site-packages/plotly/express/_chart_types.py in line(data_frame, x, y, line_group, color, line_dash, hover_name, hover_data, custom_data, text, facet_row, facet_col, facet_col_wrap, error_x, error_x_minus, error_y, error_y_minus, animation_frame, animation_group, category_orders, labels, orientation, color_discrete_sequence, color_discrete_map, line_dash_sequence, line_dash_map, log_x, log_y, range_x, range_y, line_shape, render_mode, title, template, width, height)
    242     a polyline mark in 2D space.
    243     """
--> 244     return make_figure(args=locals(), constructor=go.Scatter)
    245 
    246 

~/opt/anaconda3/envs/tsa_course/lib/python3.7/site-packages/plotly/express/_core.py in make_figure(args, constructor, trace_patch, layout_patch)
   1753     apply_default_cascade(args)
   1754 
-> 1755     args = build_dataframe(args, constructor)
   1756     if constructor in [go.Treemap, go.Sunburst] and args["path"] is not None:
   1757         args = process_dataframe_hierarchy(args)

~/opt/anaconda3/envs/tsa_course/lib/python3.7/site-packages/plotly/express/_core.py in build_dataframe(args, constructor)
   1311 
   1312     df_output, wide_id_vars = process_args_into_dataframe(
-> 1313         args, wide_mode, var_name, value_name
   1314     )
   1315 

~/opt/anaconda3/envs/tsa_course/lib/python3.7/site-packages/plotly/express/_core.py in process_args_into_dataframe(args, wide_mode, var_name, value_name)
   1117                         if argument == "index":
   1118                             err_msg += "\n To use the index, pass it in directly as `df.index`."
-> 1119                         raise ValueError(err_msg)
   1120                 elif length and len(df_input[argument]) != length:
   1121                     raise ValueError(

ValueError: Value of 'y' is not the name of a column in 'data_frame'. Expected one of ['yieldsp'] but received: Yield Spread

Before posting this question, I also looked at the solution of a similar question posted in StackOverflow, but that was not using a DateTime index, and so I was unable to resolve the error.

Answer 1

Your data sample and description of your data is incomplete. You're displaying your data as data['yieldsp'], but judging by your attempt to run px.line you've got other variables like continent in your data as well.

Anyway, what you're trying to de here is to run px.line on a dataset of wide format. And that's possible with the very latest versions of px.express. But what will not work is assigning a string to the y method and expect that you'll name your line that way. y is a method that takes data as an argument defined as a string as a reference to your dataset. Using go.Scatter() you could use name='yieldsp to rename your line. But that's not possible here. So the easiest thing here would be to rename your variable in your dataset before plotting. You still haven't provided a complete data sample, but here is how you can build your plot given that continent is not in your dataset.

Plot:

Complete code:

import plotly.express as px
import pandas as pd

# data that hopefullt represents your real world dataset
data = pd.DataFrame({'Date': {0: '1993-10-01',
                              1: '1993-10-04',
                              2: '1993-10-05',
                              3: '1993-10-06',
                              4: '1993-10-07'},
                     'yieldspd': {0: 2.36, 1: 2.32, 2: 2.29, 3: 2.31, 4: 2.28}})
data.set_index('Date', inplace = True)

# rename 'yieldspd'
data = data.rename(columns={'yieldspd': 'Yield Spread'})

# produce figure
fig = px.line(data, x = data.index, y ='Yield Spread', line_shape="spline")

# show figure
fig.show()