Inflation Forecasting for a Leading Statistical Authority in the MENA Region - Beinex
Tableau Free Trial Alteryx Free Trial Aurex Free Trial

Inflation Forecasting for a Leading Statistical Authority in the MENA Region

Industry Industry

Industry

Public Administration

Region Region

Region

UAE

AI AI

Technology

ML

Client Client

Client

Leading Statistical Authority

Requirement Requirement

Requirement

Understanding historical inflation data of the economy and forecasting the CPI for the upcoming 12 Months.​

Challenges Challenges

Challenges

  • Unavailability of data on KPIs that impact inflation
  • Data sitting in silos across different formats across databases/schemas
  • Key Factors​: Bank Rate, Private Organisations’ debt, GDP, Unemployment, PMI, Deposable Income, Labour Productivity, ERI, Commodity Index, Oil price in Abu Dhabi, ERI, Abu Dhabi Stock Exchange.​
    • Process Process

    Process

  • Identified models and multiple factors used by different countries using primary research.​
  • The correlation of the variables among themselves and eliminated any factors having a VIF above 70% ​
  • Ran recurring trials using different fields, time periods, and models to achieve the least error.​
  • FB Prophet showed more accuracy and the least error compared to the other models
    • Result Result

    Result

    A model with an accuracy of 99.2% was achieved and deployed to predict inflation for the next 12 months.

    The architectural workflow for the “CPI Prediction analysis report” is shown below:
    CPI Prediction

    Key Key

    Key Takeaway

  • Boosted Forecasting Accuracy: Machine learning models like FB Prophet can significantly improve inflation prediction compared to traditional methods.
  • Data Silos Solved: Integrating data from various sources overcomes data fragmentation, leading to more robust inflation forecasting.
  • Smarter Feature Selection: Focusing on relevant factors impacting inflation, like GDP and oil prices, optimizes model performance
    • Client Client Requirement Requirement Challenges Challenges Process Process Result Result Key Takeaway Key Takeaway