23  Reshaping Data with melt() and pivot()

23.1 Packages

import pandas as pd
import plotly.express as px

23.2 Data

Run the code below to load and define the datasets to be used in this lesson.

# Temperatures dataset
temperatures = pd.DataFrame(
    {
        "country": ["Sweden", "Denmark", "Norway"],
        "1994": [1, 2, 3],
        "1995": [3, 4, 5],
        "1996": [5, 6, 7],
    }
)

# Fuels Wide dataset
fuels_wide = pd.read_csv(
    "https://raw.githubusercontent.com/the-graph-courses/idap_book/main/data/oil_per_capita_wide.csv"
)

# Eurostat Births Wide dataset
euro_births_wide = pd.read_csv(
    "https://raw.githubusercontent.com/the-graph-courses/idap_book/main/data/euro_births_wide.csv"
)

# Contracts dataset
contracts = pd.read_csv(
    "https://raw.githubusercontent.com/the-graph-courses/idap_book/main/data/chicago_contracts_20_23.csv"
)

# Population dataset
population = pd.read_csv(
    "https://raw.githubusercontent.com/the-graph-courses/idap_book/main/data/tidyr_population.csv"
)

23.3 Introduction

Reshaping is a data manipulation technique that involves re-orienting the rows and columns of a dataset. This is often required to make data easier to analyze or understand.

In this lesson, we will cover how to effectively reshape data using pandas functions.

23.4 Learning Objectives

  • Understand what wide data format is and what long data format is.
  • Learn how to reshape wide data to long data using melt().
  • Learn how to reshape long data to wide data using pivot().

23.5 What Do “Wide” and “Long” Mean?

The terms “wide” and “long” are best understood in the context of example datasets. Let’s take a look at some now.

Imagine that you have three products for which you collect sales data over three months.

Wide Format:

Product Jan Feb Mar
A 100 120 110
B 90 95 100
C 80 85 90

Long Format:

Product Month Sales
A Jan 100
A Feb 120
A Mar 110
B Jan 90
B Feb 95
B Mar 100
C Jan 80
C Feb 85
C Mar 90

Take a minute to study the two datasets to make sure you understand the relationship between them.

In the wide dataset, each observational unit (each product) occupies only one row, and each measurement (sales in Jan, Feb, Mar) is in a separate column.

In the long dataset, on the other hand, each observational unit (each product) occupies multiple rows, with one row for each measurement.

Here’s another example with mock data, where the observational units are countries:

Long Format:

Country Year GDP
USA 2020 21433
USA 2021 22940
China 2020 14723
China 2021 17734

Wide Format:

Country GDP_2020 GDP_2021
USA 21433 22940
China 14723 17734

The examples above are both time-series datasets because the measurements are repeated across time. But the concepts of wide and long are relevant to other kinds of data too.

Consider the example below, showing the number of employees in different departments of three companies:

Wide Format:

Company HR Sales IT
A 10 20 15
B 8 25 20
C 12 18 22

Long Format:

Company Department Employees
A HR 10
A Sales 20
A IT 15
B HR 8
B Sales 25
B IT 20
C HR 12
C Sales 18
C IT 22

In the wide dataset, each observational unit (each company) occupies only one row, with the repeated measurements (number of employees in different departments) spread across multiple columns.

In the long dataset, each observational unit is spread over multiple rows.

Vocab

The observational units, sometimes called statistical units, are the primary entities or items described by the dataset.

In the first example, the observational units were products; in the second example, countries; and in the third example, companies.

Practice

23.6 Practice Q: Wide or Long?

Consider the temperatures dataset created earlier:

temperatures
country 1994 1995 1996
0 Sweden 1 3 5
1 Denmark 2 4 6
2 Norway 3 5 7

Is this data in a wide or long format?

23.7 When Should You Use Wide vs. Long Data?

The truth is, it really depends on what you want to do! The wide format is great for displaying data because it’s easy to visually compare values this way. Long data is best for some data analysis tasks, like grouping and plotting.

It is essential to know how to switch from one format to the other easily. Switching from the wide to the long format, or the other way around, is called reshaping.

23.8 From Wide to Long with melt()

To practice reshaping from a wide to a long format, we’ll consider data from Our World in Data on fossil fuel consumption per capita. You can find the data here.

Below, we view this data on fossil fuel consumption per capita:

fuels_wide
Entity Code y_1970 y_1980 y_1990 y_2000 y_2010 y_2020
0 Algeria DZA 1764.8470 3532.7976 4381.6636 3351.2180 5064.9863 4877.2680
1 Argentina ARG 11677.9680 10598.3990 7046.2485 7146.8154 7966.7827 6399.2114
2 Australia AUS 23040.4550 25007.4380 23046.9510 23976.3550 23584.3070 20332.4100
3 Austria AUT 14338.8090 19064.0920 16595.1930 18189.0920 18424.1170 14934.0650
4 Azerbaijan AZE NaN NaN 13516.0190 9119.3470 4031.9407 5615.1157
... ... ... ... ... ... ... ... ...
76 United States USA 40813.9530 42365.6500 37525.5160 37730.1600 31791.3070 26895.4770
77 Uzbekistan UZB NaN NaN 6324.8677 3197.1330 1880.1338 1859.1548
78 Venezuela VEN 11138.2210 16234.0960 12404.5570 11239.9260 14948.3070 4742.6226
79 Vietnam VNM 1757.6117 439.9465 523.2565 1280.3065 2296.7590 2927.7446
80 World OWID_WRL 7217.8340 8002.0854 7074.2583 6990.4272 6879.6110 6216.8060

81 rows × 8 columns

We observe that each observational unit (each country) occupies only one row, with the repeated measurements of fossil fuel consumption (in kilowatt-hour equivalents) spread out across multiple columns. Hence, this dataset is in a wide format.

To convert it to a long format, we can use the convenient melt function. Within melt, can define the id variables, which we do not want to reshape:

fuels_long = fuels_wide.melt(id_vars=["Entity", "Code"])
fuels_long
Entity Code variable value
0 Algeria DZA y_1970 1764.8470
1 Argentina ARG y_1970 11677.9680
2 Australia AUS y_1970 23040.4550
3 Austria AUT y_1970 14338.8090
4 Azerbaijan AZE y_1970 NaN
... ... ... ... ...
481 United States USA y_2020 26895.4770
482 Uzbekistan UZB y_2020 1859.1548
483 Venezuela VEN y_2020 4742.6226
484 Vietnam VNM y_2020 2927.7446
485 World OWID_WRL y_2020 6216.8060

486 rows × 4 columns

Very easy!

Let’s sort it so it’s easier to read:

fuels_long = fuels_long.sort_values(by=['Entity', 'variable'])
fuels_long
Entity Code variable value
0 Algeria DZA y_1970 1764.8470
81 Algeria DZA y_1980 3532.7976
162 Algeria DZA y_1990 4381.6636
243 Algeria DZA y_2000 3351.2180
324 Algeria DZA y_2010 5064.9863
... ... ... ... ...
161 World OWID_WRL y_1980 8002.0854
242 World OWID_WRL y_1990 7074.2583
323 World OWID_WRL y_2000 6990.4272
404 World OWID_WRL y_2010 6879.6110
485 World OWID_WRL y_2020 6216.8060

486 rows × 4 columns

The years are now indicated in the variable variable, and all the consumption values occupy a single variable, value. We may wish to rename the variable column to year, and the value column to oil_consumption. This can be done directly in the melt function:

fuels_long = fuels_wide.melt(
    id_vars=['Entity', 'Code'],
    var_name='year',
    value_name='oil_consumption'
).sort_values(by=['Entity', 'year'])
fuels_long
Entity Code year oil_consumption
0 Algeria DZA y_1970 1764.8470
81 Algeria DZA y_1980 3532.7976
162 Algeria DZA y_1990 4381.6636
243 Algeria DZA y_2000 3351.2180
324 Algeria DZA y_2010 5064.9863
... ... ... ... ...
161 World OWID_WRL y_1980 8002.0854
242 World OWID_WRL y_1990 7074.2583
323 World OWID_WRL y_2000 6990.4272
404 World OWID_WRL y_2010 6879.6110
485 World OWID_WRL y_2020 6216.8060

486 rows × 4 columns

You may also want to remove the y_ in front of each year. This can be achieved with a string operation.

fuels_long['year'] = fuels_long['year'].str.replace('y_', '').astype(int)
fuels_long
Entity Code year oil_consumption
0 Algeria DZA 1970 1764.8470
81 Algeria DZA 1980 3532.7976
162 Algeria DZA 1990 4381.6636
243 Algeria DZA 2000 3351.2180
324 Algeria DZA 2010 5064.9863
... ... ... ... ...
161 World OWID_WRL 1980 8002.0854
242 World OWID_WRL 1990 7074.2583
323 World OWID_WRL 2000 6990.4272
404 World OWID_WRL 2010 6879.6110
485 World OWID_WRL 2020 6216.8060

486 rows × 4 columns

Here’s what we did above:

  • Used str.replace() to remove the y_ prefix from each year.
  • Converted the year column to integers using astype(int).
  • Sorted the data by “Entity” and “year” using sort_values().
Practice

23.9 Practice Q: Temperatures to Long

Convert the temperatures dataset shown below into a long format. Your answer should have the following column names: “country”, “year”, and “avg_temp”.

# Your code here
temperatures
country 1994 1995 1996
0 Sweden 1 3 5
1 Denmark 2 4 6
2 Norway 3 5 7
Practice

23.10 Practice Q: Eurostat Births to Long

For this practice question, you will use the euro_births_wide dataset from Eurostat. It shows the annual number of births in 50 European countries:

euro_births_wide.head()
country x2015 x2016 x2017 x2018 x2019 x2020 x2021
0 Belgium 122274.0 121896.0 119690.0 118319.0 117695.0 114350.0 118349.0
1 Bulgaria 65950.0 64984.0 63955.0 62197.0 61538.0 59086.0 58678.0
2 Czechia 110764.0 112663.0 114405.0 114036.0 112231.0 110200.0 111793.0
3 Denmark 58205.0 61614.0 61397.0 61476.0 61167.0 60937.0 63473.0
4 Germany 737575.0 792141.0 784901.0 787523.0 778090.0 773144.0 795492.0

The data is in a wide format. Convert it to a long format DataFrame that has the following column names: “country”, “year”, and “births_count”.

Remove the x prefix from the year columns and convert them to integers.

23.11 Using Long Data for Analysis

Let’s see why long data is often better for analysis.

Consider again the fuels_wide dataset:

fuels_wide.head()
Entity Code y_1970 y_1980 y_1990 y_2000 y_2010 y_2020
0 Algeria DZA 1764.847 3532.7976 4381.6636 3351.2180 5064.9863 4877.2680
1 Argentina ARG 11677.968 10598.3990 7046.2485 7146.8154 7966.7827 6399.2114
2 Australia AUS 23040.455 25007.4380 23046.9510 23976.3550 23584.3070 20332.4100
3 Austria AUT 14338.809 19064.0920 16595.1930 18189.0920 18424.1170 14934.0650
4 Azerbaijan AZE NaN NaN 13516.0190 9119.3470 4031.9407 5615.1157 
fuels_long.head()
Entity Code year oil_consumption
0 Algeria DZA 1970 1764.8470
81 Algeria DZA 1980 3532.7976
162 Algeria DZA 1990 4381.6636
243 Algeria DZA 2000 3351.2180
324 Algeria DZA 2010 5064.9863

If we want to find the average fossil fuel consumption per country, this is very easy to do with the long format:

fuels_long.groupby('Entity')['oil_consumption'].mean()
Entity
Algeria           3828.796750
Argentina         8472.570833
Australia        23164.652667
Austria          16924.228000
Azerbaijan        8070.605600
                     ...     
United States    36187.010500
Uzbekistan        3315.322325
Venezuela        11784.621600
Vietnam           1537.604133
World             7063.503650
Name: oil_consumption, Length: 81, dtype: float64

But with the wide format, this is not so easy:

fuels_wide[['y_1970', 'y_1980', 'y_1990', 'y_2000', 'y_2010', 'y_2020']].mean(axis=1)
0      3828.796750
1      8472.570833
2     23164.652667
3     16924.228000
4      8070.605600
          ...     
76    36187.010500
77     3315.322325
78    11784.621600
79     1537.604133
80     7063.503650
Length: 81, dtype: float64

Imagine if you had 100 years of data!

And mean is a fairly simple operation. How would you calculate the standard deviation of fossil fuel consumption per country?

Long data is also very useful for plotting.

For example, to plot the average fossil fuel consumption per country over time, we can use the following code:

subset = fuels_long.query('Entity in ["Peru", "Iran", "China"]')
px.line(subset, x='year', y='oil_consumption', color='Entity', title='Average Fossil Fuel Consumption per Country')

To create a plot like this with the wide format is not directly possible, since the data you want to plot is scattered across multiple columns.

So as you can see, while wide data is great for display, long data is very useful for analysis and plotting.

23.12 From Long to Wide

Now you know how to reshape from wide to long with melt(). How about going the other way, from long to wide? For this, you can use the pivot() function.

But before we see how to use this function to manipulate long data, let’s first consider where you’re likely to run into long data.

While wide data tends to come from external sources (as we have seen above), long data, on the other hand, is likely to be created by you while data wrangling, especially in the course of grouped aggregations.

Let’s see an example of this now.

We will use a dataset of contracts granted by the city of Chicago from 2020 to 2023. You can find more information about the data here.

contracts
year approval_date description contract_num revision_num specification_num contract_type start_date end_date department vendor_name vendor_id address_1 address_2 city state zip award_amount procurement_type contract_pdf
0 2020 2020-01-02 LEASE 24406 32 96136 PROPERTY LEASE NaN NaN NaN 8700 BUILDING LLC 89123305A 7300 S NARRAGANSETT NaN BEDFORD PARK Illinois 60638 321.1 NaN NaN
1 2020 2020-01-03 DFSS-HHS-CS-CEL: 113798 0 1070196 DELEGATE AGENCY 12/01/2019 11/30/2022 DEPT OF FAMILY AND SUPPORT SERVICES CATHOLIC CHARITIES OF THE ARCHDIOCESE OF CHICAGO 102484615A 1 E BANKS ST NaN CHICAGO Illinois 60670 17692515.0 NaN NaN
2 2020 2020-01-03 DFSS-HHS-CS-CEL: 113819 0 1070196 DELEGATE AGENCY 12/01/2019 11/30/2022 DEPT OF FAMILY AND SUPPORT SERVICES KIMBALL DAYCARE CENTER & KINDERGARTEN INC 105458567Z 1636-1638 N KIMBALL AVE NaN CHICAGO Illinois 60647 11461500.0 NaN http://ecm.cityofchicago.org/eSMARTContracts/s...
3 2020 2020-01-03 DFSS-HHS-CS-CEL: 113818 0 1070196 DELEGATE AGENCY 12/01/2019 11/30/2022 DEPT OF FAMILY AND SUPPORT SERVICES JUDAH INTERNATIONAL OUTREACH MINISTRIES, INC 94219962X 856 N PULASKI RD NaN CHICAGO Illinois 60651 2356515.0 NaN http://ecm.cityofchicago.org/eSMARTContracts/s...
4 2020 2020-01-03 DFSS-HHS-CS-CEL: 113820 0 1070196 DELEGATE AGENCY 12/01/2019 11/30/2022 DEPT OF FAMILY AND SUPPORT SERVICES Marillac St. Vincent Family Services Inc DBA S... 97791861L 212 S FRANCISCO AVENUE EFT NaN CHICAGO Illinois 60612 3666015.0 NaN http://ecm.cityofchicago.org/eSMARTContracts/s...
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
28823 2023 2023-12-29 DFSS-CORP-HL-PSH: 220413 3 1221503 DELEGATE AGENCY NaN NaN DEPT OF FAMILY AND SUPPORT SERVICES INNER VOICE INC. 6231926M 1621 W WALNUT ST FL 1ST NaN CHICAGO Illinois 60612 0.0 NaN NaN
28824 2023 2023-12-29 DFSS-CORP-YS-OST: 253846 0 1247493 DELEGATE AGENCY NaN NaN DEPT OF FAMILY AND SUPPORT SERVICES AFTER-SCHOOL MATTERS, INC.|CLEANED-UP 72580818P 66 E RANDOLPH ST FL 1ST NaN CHICAGO Illinois 60601 32000.0 NaN NaN
28825 2023 2023-12-29 DFSS-IDHS-HL-INTHS: 253843 0 1235949 DELEGATE AGENCY NaN NaN DEPT OF FAMILY AND SUPPORT SERVICES BREAKTHROUGH URBAN MINISTRIES, INC. 94722896V 402 N ST LOUIS AVENUE EFT NaN CHICAGO Illinois 60624 14400.0 NaN NaN
28826 2023 2023-12-29 CDPH-RW-PA: ESS-HRSA PO 116685 CHICAGO HOUSE A... 192085 1 1095441 DELEGATE AGENCY NaN NaN DEPARTMENT OF HEALTH CHICAGO HOUSE & SOCIAL SERVICE AGENCY 105470138T 2229 S MICHIGAN AVE 304 EFT NaN CHICAGO Illinois 60616 -32025.2 NaN NaN
28827 2023 2023-12-29 DFSS-HHS-CS-CEL: 222199 1 1070196 DELEGATE AGENCY NaN NaN DEPT OF FAMILY AND SUPPORT SERVICES ALLISON'S INFANT & TODDLER CENTER INC 62751817Z 234 E 115TH ST FL 1ST NaN CHICAGO Illinois 60628 141923.0 NaN NaN

28828 rows × 20 columns

Each row corresponds to one contract, and we have each contract’s ID number, the year in which it was granted, the amount of the contract, and the vendor’s name and address, among other variables.

Now, consider the following grouped summary of the contracts dataset, which shows the number of contracts by state of the vendor in each year:

contracts_summary = contracts.groupby("state")["year"].value_counts().reset_index()
contracts_summary
state year count
0 Alabama 2023 7
1 Alabama 2021 2
2 Alabama 2020 1
3 Alabama 2022 1
4 Arizona 2020 3
... ... ... ...
128 Washington 2021 1
129 Wisconsin 2023 25
130 Wisconsin 2020 18
131 Wisconsin 2022 17
132 Wisconsin 2021 15

133 rows × 3 columns

The output of this grouped operation is a quintessentially “long” dataset. Each observational unit (each state) occupies multiple rows, with one row for each measurement (each year).

Now, let’s see how to convert such long data into a wide format with pivot().

The code is quite straightforward:

contracts_wide = contracts_summary.pivot(
    index="state", columns="year", values="count"
).reset_index()
contracts_wide.head()
year state 2020 2021 2022 2023
0 Alabama 1.0 2.0 1.0 7.0
1 Arizona 3.0 1.0 3.0 2.0
2 Arkansas 1.0 NaN 1.0 NaN
3 British Columbia NaN 1.0 NaN NaN
4 California 36.0 42.0 43.0 38.0

As you can see, pivot() has three important arguments:

  • index defines which column(s) to use as the new index. In our case, it’s the “state” since we want each row to represent one state.
  • columns identifies which variable to use to define column names in the wide format. In our case, it’s the “year”. You can see that the years are now the column names.
  • values specifies which values will become the core of the wide data format. In our case, it’s the number of contracts “count”.

You might also want to have the years be your primary observational units, with each year occupying one row. This can be carried out similarly to the above example, but with year as the index and state as the columns:

contracts_wide_year = contracts_summary.pivot(
    index="year", columns="state", values="count"
).reset_index()
contracts_wide_year
state year Alabama Arizona Arkansas British Columbia California Canada Colorado Connecticut Delaware ... Oregon Pennsylvania Rhode Island South Carolina Tennessee Texas Vermont Virginia Washington Wisconsin
0 2020 1.0 3.0 1.0 NaN 36.0 1.0 6.0 1.0 NaN ... 5.0 20.0 1.0 2.0 2.0 25.0 NaN 4.0 NaN 18.0
1 2021 2.0 1.0 NaN 1.0 42.0 1.0 2.0 5.0 NaN ... NaN 24.0 NaN 3.0 2.0 24.0 1.0 4.0 1.0 15.0
2 2022 1.0 3.0 1.0 NaN 43.0 1.0 7.0 3.0 1.0 ... NaN 31.0 NaN 2.0 3.0 37.0 NaN 7.0 NaN 17.0
3 2023 7.0 2.0 NaN NaN 38.0 NaN 6.0 5.0 2.0 ... NaN 37.0 NaN 1.0 3.0 28.0 NaN 9.0 NaN 25.0

4 rows × 44 columns

Here, the unique observation units (our rows) are now the years (2020, 2021, 2022, 2023).

Practice

23.13 Practice Q: Temperatures back to Wide

Convert the long temperatures_long dataset you created above back to a wide format. Your answer should have the following column names: “country”, “1994”, “1995”, and “1996”.

# Your code here
Practice

23.14 Practice Q: Population to Wide

The population dataset shows the populations of 219 countries over time.

Reshape this data into a wide format.

population
country year population
0 Afghanistan 1995 17586073
1 Afghanistan 1996 18415307
2 Afghanistan 1997 19021226
3 Afghanistan 1998 19496836
4 Afghanistan 1999 19987071
... ... ... ...
4040 Zimbabwe 2009 12888918
4041 Zimbabwe 2010 13076978
4042 Zimbabwe 2011 13358738
4043 Zimbabwe 2012 13724317
4044 Zimbabwe 2013 14149648

4045 rows × 3 columns

23.15 Reshaping Can Be Hard

We have mostly looked at very simple examples of reshaping here, but in the wild, reshaping can be difficult to do accurately.

When you run into such cases, we recommend looking at the official documentation of reshaping from the pandas team, as it is quite rich in examples.

23.16 Wrap-Up

Congratulations! You’ve gotten the hang of reshaping data with pandas.

You now understand the differences between wide and long formats and can skillfully use melt() and pivot() to transform your data as needed.

23.17 Solution Q: Wide or Long?

The data is in a wide format.

23.18 Solution Q: Temperatures to Long

# Melt the wide data into long format
temperatures_long = temperatures.melt(
    id_vars=["country"], var_name="year", value_name="avgtemp"
)

# Display the long format data
temperatures_long
country year avgtemp
0 Sweden 1994 1
1 Denmark 1994 2
2 Norway 1994 3
3 Sweden 1995 3
4 Denmark 1995 4
5 Norway 1995 5
6 Sweden 1996 5
7 Denmark 1996 6
8 Norway 1996 7

23.19 Solution Q: Eurostat Births to Long

# Melt the wide data into long format
births_long = euro_births_wide.melt(
    id_vars=["country"], var_name="year", value_name="births_count"
)

# Display the long format data
births_long
country year births_count
0 Belgium x2015 122274.0
1 Bulgaria x2015 65950.0
2 Czechia x2015 110764.0
3 Denmark x2015 58205.0
4 Germany x2015 737575.0
... ... ... ...
345 Ukraine x2021 212.0
346 Armenia x2021 271983.0
347 Azerbaijan x2021 NaN
348 Georgia x2021 112284.0
349 NaN x2021 45946.0

350 rows × 3 columns

23.20 Solution Q: Temperatures back to Wide

# Pivot the long data into wide format
temperatures_wide = temperatures_long.pivot(
    index="country", columns="year", values="avgtemp"
).reset_index()

# Display the wide format data
temperatures_wide
year country 1994 1995 1996
0 Denmark 2 4 6
1 Norway 3 5 7
2 Sweden 1 3 5

23.21 Solution Q: Population to Wide

# Pivot the long data into wide format
population_wide = population.pivot(
    index="country", columns="year", values="population"
).reset_index()

# Display the wide format data
population_wide
year country 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
0 Afghanistan 17586073.0 18415307.0 19021226.0 19496836.0 19987071.0 20595360.0 21347782.0 22202806.0 23116142.0 24018682.0 24860855.0 25631282.0 26349243.0 27032197.0 27708187.0 28397812.0 29105480.0 29824536.0 30551674.0
1 Albania 3357858.0 3341043.0 3331317.0 3325456.0 3317941.0 3304948.0 3286084.0 3263596.0 3239385.0 3216197.0 3196130.0 3179573.0 3166222.0 3156608.0 3151185.0 3150143.0 3153883.0 3162083.0 3173271.0
2 Algeria 29315463.0 29845208.0 30345466.0 30820435.0 31276295.0 31719449.0 32150198.0 32572977.0 33003442.0 33461345.0 33960903.0 34507214.0 35097043.0 35725377.0 36383302.0 37062820.0 37762962.0 38481705.0 39208194.0
3 American Samoa 52874.0 53926.0 54942.0 55899.0 56768.0 57522.0 58176.0 58729.0 59117.0 59262.0 59117.0 58652.0 57919.0 57053.0 56245.0 55636.0 55274.0 55128.0 55165.0
4 Andorra 63854.0 64274.0 64090.0 63799.0 64084.0 65399.0 68000.0 71639.0 75643.0 79060.0 81223.0 81877.0 81292.0 79969.0 78659.0 77907.0 77865.0 78360.0 79218.0
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
212 Wallis and Futuna Islands 14143.0 14221.0 14309.0 14394.0 14460.0 14497.0 14501.0 14476.0 14422.0 14344.0 14246.0 14126.0 13988.0 13840.0 13697.0 13565.0 13451.0 13353.0 13272.0
213 West Bank and Gaza Strip 2598393.0 2722497.0 2851993.0 2980563.0 3099951.0 3204572.0 3291620.0 3363542.0 3426549.0 3489743.0 3559856.0 3638829.0 3725076.0 3817551.0 3914035.0 4012880.0 4114199.0 4218771.0 4326295.0
214 Yemen 15018201.0 15578640.0 16088019.0 16564235.0 17035531.0 17522537.0 18029989.0 18551068.0 19081306.0 19612696.0 20139661.0 20661714.0 21182162.0 21703571.0 22229625.0 22763008.0 23304206.0 23852409.0 24407381.0
215 Zambia 8841338.0 9073311.0 9320089.0 9577483.0 9839179.0 10100981.0 10362137.0 10625423.0 10894519.0 11174650.0 11470022.0 11781612.0 12109620.0 12456527.0 12825031.0 13216985.0 13633796.0 14075099.0 14538640.0
216 Zimbabwe 11639364.0 11846110.0 12045813.0 12229500.0 12384727.0 12503652.0 12586763.0 12640922.0 12673103.0 12693047.0 12710589.0 12724308.0 12740160.0 12784041.0 12888918.0 13076978.0 13358738.0 13724317.0 14149648.0

217 rows × 20 columns