Government
Prediction of Singapore's Corruption Perception Index (CPI)
Mean Absolute Percentage Error (MAPE): 2%
Tags: #CorruptionIndex #CPI #Prediction #Forecasting #DataVisualisation #QlikView
Background
Corruption Perceptions Index (CPI) is a measure of government corruption by country. Index ranges from 0 to 100, with higher indices indicating lower levels of corruption. Amidst the COVID-19 pandemic, countries face multiple challenges that they had never experienced before. The pandemic has also forced large scale expenditure by governments, which may not be subjected to the thorough checks and balances that happen under normal circumstances.
Solution
Analytico Asia leveraged on data from Transparency International to create a Qlik Sense Dashboard of CPI across different countries, and predict the CPI of Singapore. Using Time-Series Analysis, we were able to achieve a forecast of very low error of margins with a Mean Absolute Percentage Error (MAPE) of 2%.
As seen in the model results below, the predicted bars in green are very close to the actual bars in purple, indicating highly accurate predictions.
Comparison of CPI between 2012 and 2021
The top 3 nations that improved in CPI between 2012 and 2021 were:
Seychelles (+18), Armenia (+15), and Italy (+14)
On the other hand, the bottom 3 nations with the highest drop in CPI were:
Saint Lucia (-15), Cyprus (-13), and Syria (-13)
Video Demo