Handling the IPCC database: reference¶

Content of the reference

  • Basics of IAMC database and the DataVar variable
  • Reference A: Selecting data
    1. Basic variable selection
    2. Specific year and adding a metadata column
    3. Multiple years, filtering on metadata column
    4. Calculations with datavars
    5. Calculations with different variables and joining multiple metadata colum
    6. Calculations and selection of metadata columns only
  • Reference B: Plotting data
    1. Basic line plot of variable over time
    2. Line plot with colors by metadata column
    3. Funnel plot (continuous error bars)
    4. Scatter plot multiple variables with facets
    5. Scatter plot with trendline
    6. Combine different plots
  • Reference C: Changing the layout of a plot
    1. Change figure size
    2. Add title
    3. Update axes (title, tickformat, ticksuffix)
    4. Style of line traces
    5. Style of scatter traces
    6. Update subplot titles
    7. Exporting figures
  • Reference D: Regional data
    1. Plotting regional data using a scatter plot
    2. Plotting on a map (choropleth)

Import packages¶

In [ ]:
from ar6_utils.importdata import import_data
from ar6_utils.datavar import DataVar
from ar6_utils.plot import CATEGORIES_COLORS, IPCC_COLORS, MACROREGIONS_GEO, line_continuous_error_bars

import numpy as np
import pandas as pd
import plotly.express as px
import plotly.io as pio
import plotly.graph_objects as go
pio.templates.default = "ipcc"
pio.renderers.default = "plotly_mimetype+notebook"

Basics of IAMC database: data and metadata¶

(todo: wat zijn de variabelen, de scenarios, wat staat er in de columns, metadata, link metadata vs database)

Overzicht van hoe je van grote bak data iets kunt doen:

  • Verschillende modelleer teams, die maken verschillende scenarios, met verschillende aannames
  • Grote bak met 3000 scenarios, maar dan eerst vetting. (visualiseren: 10 scenarios erin, 5 eruit)
  • Climate vetting voor scenarios waar je leuke dingen mee kunt (temperatuur) (referentie vettingpaper)

Regionaal? Emissies en energy use per capita,

1. The data-file:¶

2. The metadata-file:¶

3. Working with the datavar object:¶

The database and the metadata are simple "dataframes", which you can use directly to start your analysis. However, it is much more convenient to use the datavar object: it allows you to easily perform calculations between variables, select scenarios matching certain criteria, and combining it directly with metadata.

There are two steps to use the datavar object:

  1. Choose a variable you are interested in (optionally combined with a certain year or region). All calculations should be performed in the first step (adding certain variables, dividing all years by a single reference year, etc.).
  2. Select which scenarios you want to have the values for. The selection is always based on the metadata columns. This step can be used to add metadata columns, by saying that all values of that metadata column need to be selected.

In [ ]:
folder = '../../Databases/snapshots/snapshot_ar6_public_1.0/'
data_filename = 'AR6_Scenarios_Database_World_v1.0.csv'
meta_filename = 'AR6_Scenarios_Database_metadata_indicators_v1.0.xlsx'

data, meta_raw = import_data(folder, data_filename, meta_filename, extra=False)
meta = meta_raw[(meta_raw["Vetting_historical"] == "PASS") & (~meta_raw["Category"].isna())] # Only keep vetted scenarios

# Create DataVar object
datavar = DataVar(data, meta)

Importing data...
Converting to standard units...
Creating metadata...
   Importing vetting...
Finished.

Reference A: selecting data¶

Selection 1: basic variable selection¶

Select a variable, all years and no extra metadata columns:

In [ ]:
datavar("Emissions|CO2").select()
Out[ ]:
Name Year Value
0 AIM/CGE 2.0 SSP1-26 2010 35.781814
1 AIM/CGE 2.0 SSP1-26 2015 36.481576
2 AIM/CGE 2.0 SSP1-26 2020 37.181338
3 AIM/CGE 2.0 SSP1-26 2025 33.485924
4 AIM/CGE 2.0 SSP1-26 2030 29.790509
... ... ... ...
22792 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy 2080 81.678209
22793 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy 2085 82.927784
22794 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy 2090 84.177359
22795 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy 2095 82.322658
22796 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy 2100 80.467957

22797 rows × 3 columns

Selection 2: specific year and adding a metadata column¶

We now only select values in the year 2050, and add the metadata column "Category".

By specifying "all", all values are selected (no additional filtering is performed, only the metadata column is added).

In [ ]:
datavar("Emissions|CO2", 2050).select({"Category": "all"})
Out[ ]:
Category Name Value
0 C1 MESSAGEix-GLOBIOM 1.0 CD-LINKS_NPi2020_400 2.737839
1 C1 REMIND-MAgPIE 2.1-4.2 CEMICS_SSP1-1p5C-minCDR 3.175846
2 C1 IMAGE 3.2 SSP1_SPA1_19I_D_LB -0.732287
3 C1 REMIND-MAgPIE 2.1-4.2 CEMICS_SSP2-1p5C-fullCDR 0.879238
4 C1 REMIND-MAgPIE 2.1-4.2 CEMICS_SSP2-1p5C-minCDR 4.770185
... ... ... ...
1197 C8 REMIND-MAgPIE 1.7-3.0 EMF33_Baseline 64.417105
1198 C8 REMIND 2.1 R2p1_SSP5-Base 84.270954
1199 C8 IMAGE 3.2 SSP5-baseline 85.025051
1200 C8 WITCH 4.6 DISCRATE_Ref_dr5p 66.286363
1201 C8 WITCH-GLOBIOM 3.1 SSP5-Baseline 88.326223

1202 rows × 3 columns

Selection 3: Multiple years, filtering on metadata column¶

Select two years and only select those values with category C1 or C2.

Note that the columns Year and Category are now added to the output.

In [ ]:
datavar("Emissions|CO2", [2030, 2100]).select({"Category": ["C1", "C2"]})
Out[ ]:
Category Name Year Value
0 C1 AIM/CGE 2.1 CD-LINKS_NPi2020_400 2030 21.604870
1 C1 AIM/CGE 2.1 CD-LINKS_NPi2020_400 2100 -6.818765
2 C1 REMIND 2.1 R2p1_SSP5-PkBudg900 2030 25.472570
3 C1 REMIND 2.1 R2p1_SSP5-PkBudg900 2100 -12.099108
4 C1 REMIND 2.1 R2p1_SSP2-PkBudg900 2030 23.810687
... ... ... ... ...
455 C2 POLES ENGAGE EN_NPi2020_300f 2100 -15.177494
456 C2 MESSAGEix-GLOBIOM_GEI 1.0 SSP2_openres_lc_100 2030 29.414185
457 C2 MESSAGEix-GLOBIOM_GEI 1.0 SSP2_openres_lc_100 2100 -17.326515
458 C2 MESSAGEix-GLOBIOM_GEI 1.0 SSP2_openres_lc_CB400 2030 25.479194
459 C2 MESSAGEix-GLOBIOM_GEI 1.0 SSP2_openres_lc_CB400 2100 -18.331699

460 rows × 4 columns

Selection 4: Calculations with datavars¶

Perform calculations with the datavars. Any combination of variables/years can be used.

All basic arithmetic operations can be performed with datavar objects.

Note that the calculations should always happen BEFORE the .select() step.

In this example, the CO2 emissions are calculated relative to 2019.

In [ ]:
(
    datavar("Emissions|CO2") / datavar("Emissions|CO2", 2019)
).select()
Out[ ]:
Name Year Value
0 AIM/CGE 2.0 SSP1-26 2010 0.965996
1 AIM/CGE 2.0 SSP1-26 2015 0.984887
2 AIM/CGE 2.0 SSP1-26 2020 1.003778
3 AIM/CGE 2.0 SSP1-26 2025 0.904014
4 AIM/CGE 2.0 SSP1-26 2030 0.804249
... ... ... ...
22792 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy 2080 1.958800
22793 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy 2085 1.988767
22794 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy 2090 2.018735
22795 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy 2095 1.974255
22796 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy 2100 1.929776

22797 rows × 3 columns

Selection 5: Calculations with different variables and joining multiple metadata columns¶

Different variables can also be used in calculations. And more meta columns can be used.

In [ ]:
(
    datavar("Primary Energy|Wind") + datavar("Primary Energy|Solar")
).select({
    "Category": ["C1", "C2"],
    "Year of peak GHG Emissions": "all",
    "Cumulative net CO2 (2020-2100, Gt CO2) (Harm-Infilled)": "all",
})
Out[ ]:
Category Year of peak GHG Emissions Cumulative net CO2 (2020-2100, Gt CO2) (Harm-Infilled) Name Year Value
0 C1 2015.0 -316.756374 MESSAGE-GLOBIOM 1.0 EMF33_1.5C_full 2010 4.182713
1 C1 2015.0 -316.756374 MESSAGE-GLOBIOM 1.0 EMF33_1.5C_full 2015 11.409575
2 C1 2015.0 -316.756374 MESSAGE-GLOBIOM 1.0 EMF33_1.5C_full 2020 18.636437
3 C1 2015.0 -316.756374 MESSAGE-GLOBIOM 1.0 EMF33_1.5C_full 2025 35.580142
4 C1 2015.0 -316.756374 MESSAGE-GLOBIOM 1.0 EMF33_1.5C_full 2030 52.523846
... ... ... ... ... ... ...
3365 C2 2025.0 547.363058 REMIND-MAgPIE 2.1-4.2 EN_INDCi2030_600f_NDCp 2080 578.611000
3366 C2 2025.0 547.363058 REMIND-MAgPIE 2.1-4.2 EN_INDCi2030_600f_NDCp 2085 655.113850
3367 C2 2025.0 547.363058 REMIND-MAgPIE 2.1-4.2 EN_INDCi2030_600f_NDCp 2090 731.616700
3368 C2 2025.0 547.363058 REMIND-MAgPIE 2.1-4.2 EN_INDCi2030_600f_NDCp 2095 808.748550
3369 C2 2025.0 547.363058 REMIND-MAgPIE 2.1-4.2 EN_INDCi2030_600f_NDCp 2100 885.880400

3370 rows × 6 columns

Selection 6: Calculations and selection of metadata columns only¶

Until now, the datavar object was only used to get variable data and to perform calculations with those. The .select(...) command was used to join metadata columns to that variable data, and potentially do some filtering. However, it often happens that you only want to get metadata, without any variable (for example, when you want to plot cumulative emissions as function of net-zero year). This is possible using datavar(meta=...):

In [ ]:
selection6 = datavar(
    meta=[
        "Cumulative net CO2 (2020-2100, Gt CO2) (Harm-Infilled)",
        "Year of netzero CO2 emissions (Harm-Infilled) table",
    ]
).select({"Category": "all"})
selection6.head()
Out[ ]:
Category Name Cumulative net CO2 (2020-2100, Gt CO2) (Harm-Infilled) Year of netzero CO2 emissions (Harm-Infilled) table
0 C1 MESSAGEix-GLOBIOM 1.0 CD-LINKS_NPi2020_400 -110.696287 2052.0
1 C1 REMIND-MAgPIE 2.1-4.2 CEMICS_SSP1-1p5C-minCDR 513.769205 2068.0
2 C1 IMAGE 3.2 SSP1_SPA1_19I_D_LB 308.957896 2049.0
3 C1 REMIND-MAgPIE 2.1-4.2 CEMICS_SSP2-1p5C-fullCDR 431.417534 2052.0
4 C1 REMIND-MAgPIE 2.1-4.2 CEMICS_SSP2-1p5C-minCDR 520.569379 2069.0

Note that we now still join the Category column in the .select({"Category": "all"}). The same could be achieved by simply adding Category to the list of metadata columns in datavar. The only difference is that the dataframe is not sorted anymore by category.

In [ ]:
selection6b = datavar(
    meta=[
        "Cumulative net CO2 (2020-2100, Gt CO2) (Harm-Infilled)",
        "Year of netzero CO2 emissions (Harm-Infilled) table",
        "Category",
    ]
).select()
selection6b.head()
Out[ ]:
Name Cumulative net CO2 (2020-2100, Gt CO2) (Harm-Infilled) Year of netzero CO2 emissions (Harm-Infilled) table Category
0 AIM/CGE 2.0 SSP1-26 942.485543 2090.0 C3
1 AIM/CGE 2.0 SSP1-34 1774.370588 NaN C5
2 AIM/CGE 2.0 SSP1-45 2473.615376 NaN C6
3 AIM/CGE 2.0 SSP1-Baseline 2994.686677 NaN C7
4 AIM/CGE 2.0 SSP4-26 926.934351 2077.0 C3

Reference B: plotting data¶

Plot 1: basic line plot of variable over time¶

In [ ]:
selection1 = datavar("Emissions|CO2").select()
selection1.head()
Out[ ]:
Name Year Value
0 AIM/CGE 2.0 SSP1-26 2010 35.781814
1 AIM/CGE 2.0 SSP1-26 2015 36.481576
2 AIM/CGE 2.0 SSP1-26 2020 37.181338
3 AIM/CGE 2.0 SSP1-26 2025 33.485924
4 AIM/CGE 2.0 SSP1-26 2030 29.790509
In [ ]:
fig1 = px.line(
    selection1,
    x="Year",
    y="Value",
    line_group="Name", # This is necessary to know which points belong on one line. In this case, all the points with the same "Name" column values.
)
## Note: all keywords here (`Year``, `Values`, `Name`) correspond to columns in `selection1`

fig1.show(renderer='svg') # NOTE: remove `renderer='svg'` to make the figure interactive
20202040206020802100−20 0 20 40 60 80 100 120 YearValue

Plot 2: line plot with colors by metadata column¶

In [ ]:
selection2 = datavar("Emissions|CO2").select({"Category": "all"})
selection2.head()
Out[ ]:
Category Name Year Value
0 C1 MESSAGEix-GLOBIOM 1.0 CD-LINKS_NPi2020_400 2010 38.542018
1 C1 MESSAGEix-GLOBIOM 1.0 CD-LINKS_NPi2020_400 2015 39.078620
2 C1 MESSAGEix-GLOBIOM 1.0 CD-LINKS_NPi2020_400 2020 39.615223
3 C1 MESSAGEix-GLOBIOM 1.0 CD-LINKS_NPi2020_400 2025 31.719723
4 C1 MESSAGEix-GLOBIOM 1.0 CD-LINKS_NPi2020_400 2030 23.824223
In [ ]:
fig2 = px.line(
    selection2,
    x="Year",
    y="Value",
    line_group="Name",
    color="Category",
)
fig2.show(renderer='svg') # NOTE: remove `renderer='svg'` to make the figure interactive
20202040206020802100−20 0 20 40 60 80 100 120 CategoryC1C2C3C4C5C6C7C8YearValue

Plot 3: funnel plot (continuous error bars)¶

This is not a standard feature of Plotly. Therefore, use the function line_continuous_error_bars from ar6_utils.plot. It uses the exact same syntax as the default px.line from Plotly Express, but needs an extra argument groupby_columns: a list of columns in the selection that should be grouped together. Typically, this would be Year and Category (or another metadata column).

Default quantiles are 0.05 and 0.95, which can be changed using q_low and q_high arguments.

In [ ]:
selection3 = datavar("Emissions|CO2").select({"Category": "all"})
selection3.head()
Out[ ]:
Category Name Year Value
0 C1 MESSAGEix-GLOBIOM 1.0 CD-LINKS_NPi2020_400 2010 38.542018
1 C1 MESSAGEix-GLOBIOM 1.0 CD-LINKS_NPi2020_400 2015 39.078620
2 C1 MESSAGEix-GLOBIOM 1.0 CD-LINKS_NPi2020_400 2020 39.615223
3 C1 MESSAGEix-GLOBIOM 1.0 CD-LINKS_NPi2020_400 2025 31.719723
4 C1 MESSAGEix-GLOBIOM 1.0 CD-LINKS_NPi2020_400 2030 23.824223
In [ ]:
fig3 = line_continuous_error_bars(
    selection3,
    ["Year", "Category"], # these are the groupby_columns,
    x="Year",
    y="Value",
    color="Category",
)
fig3.show()

Plot 4: Scatter plot multiple variables with facets¶

In [ ]:
selection4 = datavar(
    meta=[
        "Year of netzero CO2 emissions (Harm-Infilled) table",
        "Year of netzero GHG emissions (Harm-Infilled) table",
        "Cumulative net CO2 (2020 to netzero, Gt CO2) (Harm-Infilled)"
    ]
).select({"Category": ["C1", "C2", "C3", "C4", "C5"]})

selection4.head()
Out[ ]:
Category Name Year of netzero CO2 emissions (Harm-Infilled) table Year of netzero GHG emissions (Harm-Infilled) table Cumulative net CO2 (2020 to netzero, Gt CO2) (Harm-Infilled)
0 C1 WITCH-GLOBIOM 4.4 CD-LINKS_NPi2020_400 2055.0 2069.0 473.416283
1 C1 GCAM 5.3 R_MAC_50_n8 2045.0 inf 512.002256
2 C1 REMIND-MAgPIE 2.1-4.2 CEMICS_SSP2-1p5C-minCDR 2069.0 inf 528.644620
3 C1 REMIND-MAgPIE 2.1-4.2 CEMICS_SSP2-1p5C-fullCDR 2052.0 inf 539.559537
4 C1 REMIND-MAgPIE 2.1-4.2 CEMICS_SSP1-1p5C-minCDR 2068.0 inf 537.313335
In [ ]:
fig4 = px.scatter(
    selection4,
    x="Cumulative net CO2 (2020 to netzero, Gt CO2) (Harm-Infilled)",
    y=[
        "Year of netzero CO2 emissions (Harm-Infilled) table",
        "Year of netzero GHG emissions (Harm-Infilled) table",
    ],
    color="Category",
    facet_col="variable",
)
fig4.show()

Plot 5: Scatter plot with trendline¶

When creating a scatter plot, you can easily add a linear fit (trendline) to the data. By default this happens for each trace (for each color group). To create a global trendline, add the argument trendline_scope="overall".

In the plot, hover over the trendline to get the equation of the trendline.

In [ ]:
selection5 = datavar(
    meta=[
        "Year of netzero CO2 emissions (Harm-Infilled) table",
        "Median warming in 2100 (FaIRv1.6.2)",
    ]
).select({"Category": "all"})
selection5.head()
Out[ ]:
Category Name Year of netzero CO2 emissions (Harm-Infilled) table Median warming in 2100 (FaIRv1.6.2)
0 C1 MESSAGEix-GLOBIOM 1.0 CD-LINKS_NPi2020_400 2052.0 1.095274
1 C1 REMIND-MAgPIE 2.1-4.2 CEMICS_SSP1-1p5C-minCDR 2068.0 1.274399
2 C1 IMAGE 3.2 SSP1_SPA1_19I_D_LB 2049.0 1.276742
3 C1 REMIND-MAgPIE 2.1-4.2 CEMICS_SSP2-1p5C-fullCDR 2052.0 1.383032
4 C1 REMIND-MAgPIE 2.1-4.2 CEMICS_SSP2-1p5C-minCDR 2069.0 1.415305
In [ ]:
fig5 = px.scatter(
    selection5,
    x="Year of netzero CO2 emissions (Harm-Infilled) table",
    y="Median warming in 2100 (FaIRv1.6.2)",
    color="Category",
    trendline="ols",
    trendline_scope="overall", # Remove this line to create trendline by category
)
fig5.show()

Plot 6: Combine different plots¶

It is possible to combine different plots: for example, a funnel graph and a line plot for specific scenarios, or a scatter plot of all scenarios combined with a scatter plot of Illustrative Mitigation Pathways.

To combine them, first make the figures separately. Then, create a new empty figure which uses the traces (data) of both figures.

In [ ]:
fig1 = line_continuous_error_bars(
    datavar("Emissions|CO2").select({"Category": "all"}),
    ["Year", "Category"],
    x="Year",
    y="Value",
    color="Category",
    with_median=False,
)

fig2 = px.line(
    datavar("Emissions|CO2").select({"IMP_marker": "all"}),
    x="Year",
    y="Value",
    color="IMP_marker",
    color_discrete_sequence=IPCC_COLORS,
)

fig_combined = go.Figure(data=fig1.data + fig2.data, layout=fig1.layout)
fig_combined

Reference C: changing the layout of a plot¶

This section contains some sample code on how to change the layout of a Plotly figure. A full reference is available here: https://plotly.com/python/reference/layout/.

In Plotly, the layout (styling) of a figure is changed after creating the figure. The layout is changed by calling the function .update_layout() on the figure:

fig = px.scatter(...)   # Create the figure
fig.update_layout(      # Update the layout
    ...
)
In [ ]:
# Example figure:
fig = px.line(
    datavar("Emissions|CO2").select({"Category": ["C1", "C2"]}),
    x="Year",
    y="Value",
    color="Category",
    line_group="Name",
)
fig.show(renderer='svg') # NOTE: remove `renderer='svg'` to make the figure interactive
20202040206020802100−20 −10 0 10 20 30 40 50 CategoryC1C2YearValue

Layout 1: Change figure size¶

In [ ]:
fig.update_layout(
    width=500,
    height=300,
)

fig.show(renderer='svg') # NOTE: remove `renderer='svg'` to make the figure interactive
20202040206020802100−20 0 20 40 CategoryC1C2YearValue

Layout 2: Add title¶

In [ ]:
fig.update_layout(
    title="CO2 emissions by climate category"
)

fig.show(renderer='svg') # NOTE: remove `renderer='svg'` to make the figure interactive
20202040206020802100−20 0 20 40 CategoryC1C2CO2 emissions by climate categoryYearValue

Layout 3: Update axes¶

Instead of updating the general layout of the figure using fig.update_layout(...), the axes can be updated using fig.update_xaxes(...) and fig.update_yaxes(...). The full documentation about this can be found at

  • https://plotly.com/python/axes/
  • https://plotly.com/python/reference/layout/xaxis/
  • https://plotly.com/python/reference/layout/yaxis/
In [ ]:
fig.update_xaxes(
    range=[2020, 2050]
)

fig.update_yaxes(
    title="CO<sub>2</sub> emissions", # Note that HTML commands like <sub>2</sub> can be used for subscript font.
    ticksuffix=" GtCO<sub>2</sub>",
)

fig.show(renderer='svg') # NOTE: remove `renderer='svg'` to make the figure interactive
2020203020402050−20 GtCO​2​0 GtCO​2​20 GtCO​2​40 GtCO​2​CategoryC1C2CO2 emissions by climate categoryYearCO​2​ emissions

Layout 4: Style of line traces¶

Just like the general layout (fig.update_layout(...)) and axes (fig.update_xaxes(...)), the traces (the data itself) can be styled using fig.update_traces(...).

In [ ]:
fig.update_traces(
    line={
        'width': 1,
        'dash': 'dash',
    },
    opacity=0.75,
)

fig.show(renderer='svg') # NOTE: remove `renderer='svg'` to make the figure interactive
2020203020402050−20 GtCO​2​0 GtCO​2​20 GtCO​2​40 GtCO​2​CategoryC1C2CO2 emissions by climate categoryYearCO​2​ emissions

Layout 5: Styling of scatter traces¶

A scatter plot can also be styled using fig.update_traces(...) but using marker=.... In this example we start with the example from Plot 4.

In [ ]:
# First we update only the left column using `col=1`:
fig4.update_traces(
    col=1,
    marker={
        "size": 4,
        "symbol": "star",
    },
)

# Then we update the right column using `col=2`:
fig4.update_traces(
    col=2,
    marker={
        "symbol": "diamond",
        "size": 10
    }
)

Layout 6: Update subplot titles¶

Starting with the previous figure (from Plot 4 and Layout 5), we see that the subplots automatically get titles that start with "variable=". You can remove this by updating the subplots titles:

In [ ]:
fig4.for_each_annotation(lambda a: a.update(text=a.text.split("=")[-1]))

fig4.show()

Layout 7: Exporting figures¶

After having created and styled a figure, you can export it to various formats: png, svg, pdf and more. The advantage of svg is that you can edit the exported SVG figure in programs like Inkscape or Adobe Illustrator to further style the figure.

In [ ]:
fig.write_image("yourfilename.png")
fig.write_image("yourfilename.svg")
fig.write_image("yourfilename.pdf")

Reference D: regional data¶

By default, when importing the data you only import the global data file and the metadata file. However, the metadata can also be combined with a regional version of the IPCC AR6 database. Here, we use the R5 regions, but the data is also available on the IIASA website on different resolutions. The DataVar object is also able to work with regional data: it simply adds an extra column Region. Import it using:

In [ ]:
data_regional_filename = 'AR6_Scenarios_Database_R5_regions_v1.0.csv'

# Don't forget the argument `only_world=False`
data_regional, _ = import_data(folder, data_regional_filename, meta_filename, extra=False, onlyworld=False)

# Create DataVar object. Don't foget the argument `is_regional=True`
datavar_regional = DataVar(data_regional, meta, is_regional=True)

Importing data...
Converting to standard units...
Creating metadata...
   Importing vetting...
Finished.

Regional 1: Plotting regional data using a scatter plot¶

In [ ]:
selection_regional = datavar_regional(["Emissions|CO2", "Emissions|CH4"], 2100).select()
selection_regional
Out[ ]:
Name Region Variable Value
0 AIM/CGE 2.0 SSP1-26 R5ASIA Emissions|CH4 32.958800
1 AIM/CGE 2.0 SSP1-26 R5ASIA Emissions|CO2 -0.108355
2 AIM/CGE 2.0 SSP1-26 R5LAM Emissions|CH4 14.563300
3 AIM/CGE 2.0 SSP1-26 R5LAM Emissions|CO2 -0.543469
4 AIM/CGE 2.0 SSP1-26 R5MAF Emissions|CH4 39.010000
... ... ... ... ...
11919 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy R5MAF Emissions|CO2 14.810541
11920 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy R5OECD90+EU Emissions|CH4 45.626720
11921 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy R5OECD90+EU Emissions|CO2 12.973879
11922 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy R5REF Emissions|CH4 57.687099
11923 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy R5REF Emissions|CO2 6.321122

11924 rows × 4 columns

In [ ]:
# To plot CH4 emissions as function of CO2 emissions,
# we need a column CO2 and a column CH4, instead of a column Variable and a column Value.
# We call this going from long format to wide format.
selection_regional_wide = (
    selection_regional
    .set_index(["Name", "Region", "Variable"])  # Create an index of all columns except Value
    ["Value"]                                   # Select the only remaining column Value
    .unstack("Variable")                        # Unstack the variable we want to make wide
    .reset_index()                              # Go back to normal dataframe
)
selection_regional_wide
Out[ ]:
Variable Name Region Emissions|CH4 Emissions|CO2
0 AIM/CGE 2.0 SSP1-26 R5ASIA 32.958800 -0.108355
1 AIM/CGE 2.0 SSP1-26 R5LAM 14.563300 -0.543469
2 AIM/CGE 2.0 SSP1-26 R5MAF 39.010000 -0.052376
3 AIM/CGE 2.0 SSP1-26 R5OECD90+EU 22.528100 0.544884
4 AIM/CGE 2.0 SSP1-26 R5REF 3.632400 -0.192054
... ... ... ... ...
5957 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy R5ASIA 131.380225 33.816931
5958 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy R5LAM 57.546215 6.469990
5959 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy R5MAF 83.388491 14.810541
5960 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy R5OECD90+EU 45.626720 12.973879
5961 WITCH-GLOBIOM 4.4 CD-LINKS_NoPolicy R5REF 57.687099 6.321122

5962 rows × 4 columns

In [ ]:
fig = px.scatter(
    selection_regional_wide,
    x="Emissions|CO2",
    y="Emissions|CH4",
    color="Region",
    hover_name="Name",
)
fig

Regional 2: Plotting on a map (choropleth)¶

A nice visual way of plotting regional data is on a map. This is called a choropleth plot. However, a map can only display one-dimensional data: one value per region. We can either select one single scenario for this, or we can aggregate all scenarios in a selection using the following aggregations:

  • mean()
  • median()
  • quantile(0.05) (or any other quantile)

To achieve this we need to group the selection by all relevant variables (Region and possibly Year and/or Variable, and all metadata columns like Category if you included it in .select()):

In [ ]:
selection = datavar_regional("Emissions|CO2", [2020, 2050, 2100]).select({"Category": ["C1", "C3", "C7"]})
selection_grouped = selection.groupby(["Region", "Year", "Category"])["Value"].median().reset_index()
selection_grouped.head()
Out[ ]:
Region Year Category Value
0 R5ASIA 2020 C1 17.393373
1 R5ASIA 2020 C3 17.090441
2 R5ASIA 2020 C7 18.239535
3 R5ASIA 2050 C1 1.260943
4 R5ASIA 2050 C3 5.077181
In [ ]:
fig_choropleth = px.choropleth(
    selection_grouped,
    locations="Region",
    geojson=MACROREGIONS_GEO,   # Regional definitions are given in this variable
    color="Value",
    facet_col="Year",           # Ignore if you don't have different years
    facet_row="Category",       # Ignore if you don't differentiate by category
    projection="natural earth",
    labels={"Value": "CO<sub>2</sub> emissions"}, # Not necessary, it just gives the colorbar a nice title
)
fig_choropleth.update_layout(height=600)