How to Build a Global Inflation Dashboard in Python (Data Visualization Tutorial)
Raksina Samasiri
Developer Advocate Developer Advocate

Introduction

Inflation charts are everywhere - economic reports, dashboards, and news updates.
But even with all that data, I kept running into the same problem: It’s surprisingly hard to compare how inflation is evolving across multiple countries over time.

Line charts become cluttered with too many series. Tables provide precision but lack intuitive insight. So instead of searching for the “perfect dashboard,” I decided to build one.

In this article, I’ll walk through how to use LSEG Datastream with Python to create a global inflation dashboard using:

  • Line charts (to understand trends)
  • Heatmaps (to identify patterns)

Prerequisites

Before getting started, make sure you have the following:

  1. Python Environment: Python 3.8 or above (I'm using Python 3.12.12)
  2. Required Libraries: Install the required Python packages
    1. DatastreamPy==2.0.30
    2. pandas==3.0.3
    3. seaborn==0.13.2
    4. matplotlib==3.10.9
  3. Access to LSEG Datastream: You need valid Datastream credentials to run the examples.
  4. Basic Knowledge: Familiarity with basic scripting in Python

How to Use Datastream Navigator (Quick Guide)

Here’s a simple workflow based on the screenshots below.

  1. Search for a Concept: In the search bar, enter a keyword like "inflation rate yoy". This returns multiple matching datasets.
  2. Apply Filters: On the left panel, remove irrelevant datasets quickly, narrow your results to:
    1. Category -> Economics
    2. Country or Region -> Region
    3. Activity -> Active
    4. Key Indicator -> Yes
  3. Inspect a Dataset: Click on a result to view its details.
    From the image: These tell you that this is inflation that is expressed in percentage and updated regularly
    1. Name: CPI (%YoY)
    2. Symbol: CNCPANNL
    3. Unit: Percentage
    4. Frequency: Monthly

4. Identify Naming Patterns: Looking at multiple results:

    • USCPANNL -> United States
    • UKCPANNL -> United Kingdom
    • JPCPANNL -> Japan
    • CNCPANNL -> Canada
  • Which you can infer a pattern: <Country Code> + CPANNL

5. Scale Your Dataset: Once you identify the pattern, you can build a dataset quickly. No need to manually search each country individually.

Important Note on Frequency

Although the Navigator shows Monthly Frequency . In practice, some macroeconomic datasets may behave like bi‑monthly observations (i.e. values not updated every single month), which means that the data points may appear every ~2 months. For visualization, treat it as time series (chronological order matters) and adjust display density as needed (e.g. every 4 months)

Step 1: Import Libraries Connect to DataStream

Import necessary Python library and initialize the Datastream client to query the data.

    	
            
# Step 1: Import libraries and connect to DataStream


import DatastreamPy as dsws


import matplotlib.pyplot as plt


import seaborn as sns


import pandas as pd


 


ds=dsws.DataClient(username='YOUR_USERNAME',password='YOUR_PASSWORD')

Step 2: Define Countries

Define a list of inflation series across developed and emerging markets. Here we include a mix of developed and emerging markets:

    	
            
# Step 2: Define Countries


series = [


    "USCPANNL",  # United States


    "UKCPANNL",  # United Kingdom


    "JPCPANNL",  # Japan


    "FRCPANNL",  # France


    "ITCPANNL",  # Italy


    "DECPANNL",  # Germany


    "CNCPANNL",  # Canada


    "IDCPANNL",  # Indonesia


    "INCPANNL",  # India


    "NZCPANNL"   # New Zealand


]


 


tickers_input = ",".join(series)


tickers_input

Step 3: Fetch Data

Retrieve time series data (bi‑monthly inflation observations). Datastream returns data in a wide format, where:

  • rows = time (bi-monthly)
  • columns = instruments
    	
            
# Step 3: Fetch Data


data = ds.get_data(


    tickers=tickers_input,


    start="2020-01-01",


    end="2026-01-01",


    freq="M"


)


data.head()

Step 4: Clean and Standardize

Before visualization, we clean the dataset once as below - and reuse it everywhere.

  • Convert index to datetime
  • Sort by time
  • Flatten column structure (returned data is in tuples for this ticker)
  • To make the output readable, replace technical mnemonics with readable country names.
    	
            
# Step 4: Clean and Standardize


df = data.copy()


 


# ensure datetime index


df.index = pd.to_datetime(df.index)


 


# sort by date


df = df.sort_index()


 


# FIX: columns from tuple to ticker string


df.columns = [c[0] if isinstance(c, tuple) else c for c in df.columns]


 


# map country names


country_map = {


    "USCPANNL": "US",


    "UKCPANNL": "UK",


    "JPCPANNL": "Japan",


    "FRCPANNL": "France",


    "ITCPANNL": "Italy",


    "DECPANNL": "Germany",


    "CNCPANNL": "Canada",


    "IDCPANNL": "Indonesia",


    "INCPANNL": "India",


    "NZCPANNL": "New Zealand"


}


 


df.columns = [country_map.get(c, c) for c in df.columns]


 


df.head()

Step 5: Visualizing Trends with Line Charts

Line chart helps identify global trends over time.

    	
            
# Step 5: Visualizing Trends with Line Charts


df.plot(figsize=(18, 6))


 


plt.title("Global Inflation Trends (YoY %)")


plt.xlabel("Date")


plt.ylabel("Inflation (%)")


 


plt.legend(bbox_to_anchor=(1, 1), loc='upper left')


plt.grid(alpha=0.3)


 


plt.tight_layout()


plt.show()

What this shows?

  • A global inflation surge (2021–2022)
  • A cooling trend (2023 onward)
  • Differences across economies:
    • Japan -> low inflation
    • India -> persistently higher inflation

Step 6: Reusable Heatmap Functions

To make the visualization flexible, we separate:

  • data preparation
  • visualization

Step 6.1: Prepare Heatmap Data

  • Transpose data (countries -> rows)
  • Adjust frequency (bi‑monthly vs every 4 months)
  • Format time labels
    	
            
# Step 6.1: Prepare Heatmap Data


def prepare_heatmap_data(df, freq="bi-monthly"):


    heatmap_df = df.T.copy()


 


    # convert datetime


    heatmap_df.columns = pd.to_datetime(heatmap_df.columns)


 


    if freq == "reduced":


        heatmap_df = heatmap_df.iloc[:, ::2]  # every 4 months


 


    # format date


    heatmap_df.columns = heatmap_df.columns.strftime("%Y-%m")


 


    return heatmap_df

Step 6.2: Plot Heatmap

Heatmaps reveal cross-country patterns more clearly than line charts.

    	
            
# Step 6.2: Plot Heatmap


def plot_inflation_heatmap(df, name):


    plt.figure(figsize=(16, 6))


    


    sns.heatmap(


        df,


        cmap="coolwarm",


        linewidths=0.1,


        cbar_kws={"label": "Inflation (%)"}


    )


    


    plt.title(name)


    plt.xlabel("Time")


    plt.ylabel("Country")


    


    plt.tight_layout()


    plt.show()

Step 6.3: Bi-Monthly vs Reduced Views

  • Bi‑monthly: detailed view
  • Every 4 months: cleaner comparison
    	
            
# monthly


hm_monthly = prepare_heatmap_data(df, freq="bi-monthly")


plot_inflation_heatmap(hm_monthly, "Global Inflation Heatmap (Bi-Monthly)")


        
        
    
    	
            
# quarterly


hm_quarterly = prepare_heatmap_data(df, freq="reduced")


plot_inflation_heatmap(hm_quarterly, "Global Inflation Heatmap (Every 4 Month)")

Why This Approach Works

A few key design decisions make this workflow effective:

  1. Use the right metric: Avoid unnecessary transformations by using inflation YoY directly
  2. Clean once, reuse everywhere: Mapping and formatting done early -> simpler downstream code
  3. Separate transformation from visualization: This enables flexible experimentation and reusable plotting functions
  4. Choose the right visualization
    1. Line charts -> trend over time
    2. Heatmaps -> cross-country comparison

Key Insights

From these visualizations:

  • Inflation peaked globally around 2022
  • Developed economies show similar patterns
  • Emerging markets exhibit:
    • higher baseline inflation
    • slower normalization

Takeaways for Developers

If you're building data visualizations:

  • Think about data structure first
  • Choose the right metric, not just available data
  • Clean once, reuse across visualizations
  • Match the chart type to the question

Conclusion

This project wasn’t just about building a dashboard. It was about making global data easier to understand. Often, the biggest improvement in visualization doesn’t come from better charts, but from better thinking about the data.

  • Register or Log in to applaud this article
  • Let the author know how much this article helped you
Sign In
Register Now
If you require assistance, please contact us here

Request Free Trial

Help & Support

Already a customer?

Get help through MyAccount

Office locations

Contact LSEG near you

Find your nearest LSEG office