Plotting SDMX with PANDAS or plot

Question

What's the easiest way to plot SDMX data with PANDAS or Plotly?

I have the following code:

import pandasdmx as sdmx 
import plotly.express as px

df = sdmx.Request('OECD').data(
  resource_id='MEI_FIN',
  key='IR3TIB.GBR+USA.M',
  params={'startTime': '1900-06', 'dimensionAtObservation': 'TimeDimension'},
).write().reset_index()
df

i end up getting the following error when trying to plot

fig = px.line(df, x="TIME_PERIOD", y='', title='Life expectancy in Country: Denmark')
fig.show()

as the following:

ValueError: Value of 'y' is not the name of a column in 'data_frame'. Expected one of `[('TIME_PERIOD', '', ''), ('IR3TIB', 'GBR', 'M'), ('IR3TIB', 'USA', 'M')] but received:`

I am pretty new with python so i would appreciate every comment that could help me with this.

score 0 · Accepted Answer · answered Jul 08 '20 at 15:26

0

I think that your main problem is due to the fact that your df is with multiindex. I'm not sure if this is what you what to achieve but you can try the following code:

import pandasdmx as sdmx 
import plotly.express as px

df = sdmx.Request('OECD').data(
  resource_id='MEI_FIN',
  key='IR3TIB.GBR+USA.M',
  params={'startTime': '1900-06', 'dimensionAtObservation': 'TimeDimension'},
).write().reset_index()

# with this we get rid of multi-index
# you could use a loop if you prefer I used
# list of comprehension

df.columns = ["_".join([c for c in col if c!='']) 
              for col in df.columns]

fig = px.line(df,
              x="TIME_PERIOD",
              y=['IR3TIB_GBR_M', 'IR3TIB_USA_M'],
              title='Life expectancy in GBR and USA')\
        .update_layout(title_x=0.5)
fig.show()

answered Jul 08 '20 at 15:26

rpanai

12,515
2
42
64

thank you so much! So how can this be implemented in dash app with a dropdown with the two countries? I've tried this with app.callback but i don't think i'm near a solution. – nonamednono Jul 09 '20 at 18:55
You can do so with a plotly only solution too. In theory, as per SO rules, you should ask another question. But the most likely thing is that you only need to adapt this [answer](https://stackoverflow.com/a/61754644/4819376) – rpanai Jul 09 '20 at 19:50
Then let me know if you have any problem with it. And please consider to upvote this answer (and the other) if it was useful to you. – rpanai Jul 09 '20 at 19:51

Plotting SDMX with PANDAS or plot

1 Answers1