Time Series Analysis

 

Predictive Analytics Using ARIMA: Forecasting Running Cost of Electric, CNG, Petrol, and Diesel Vehicles (2025–2030)

A Time-Series Forecasting Case Study for Cost-Effective Vehicle Ownership


Introduction

As fuel and energy prices fluctuate every year, choosing the most cost-effective vehicle type—Electric (EV), CNG, Petrol, or Diesel—requires more than static comparison. Predictive analytics, particularly Time-Series Forecasting using ARIMA models (Auto-Regressive Integrated Moving Average), can provide powerful insights into future running costs.

This post explores how ARIMA can forecast the cost per kilometer (₹/km) for different fuel types from 2025 to 2030, helping buyers and policy planners make data-driven decisions.


Understanding the ARIMA Model

The ARIMA (p, d, q) model is a widely used time-series forecasting technique:

  • p (Auto-regressive order): how past values influence the current value.

  • d (Differencing order): how many times the data is differenced to achieve stationarity.

  • q (Moving average order): how past forecast errors influence future values.

ARIMA works best when:

  • Historical data is available for each fuel’s running cost per km.

  • The trend is smooth, with minor random fluctuations.


Dataset Preparation (Base Year 2025)

Vehicle TypeEnergy EfficiencyEnergy PriceMaintenanceTotal Cost (₹/km, 2025)
Electric6 km/kWh₹8/kWh₹6,000/year1.73
CNG25 km/kg₹75/kg₹10,000/year3.67
Petrol15 km/l₹110/l₹12,000/year8.13
Diesel20 km/l₹100/l₹15,000/year6.00

Step 1: Generating Time-Series Data

The running cost for each fuel type (2025–2030) is treated as a separate time series.
We assume annual energy and maintenance inflation rates:

Fuel TypeAnnual Inflation RateMaintenance Growth
Electricity+3%+4%
CNG+5%+4%
Petrol+6%+4%
Diesel+5%+4%

These give us a six-year dataset for each vehicle type.


Step 2: ARIMA Model Fitting (Conceptual)

Let’s take Electric Vehicle (EV) data as an example:

Time-Series (2025–2030)

YearEV Cost (₹/km)
20251.73
20261.78
20271.83
20281.88
20291.93
20301.98

Using Python or R, we can model this data as:

from statsmodels.tsa.arima.model import ARIMA model = ARIMA(ev_data, order=(1,1,0)) results = model.fit() forecast = results.forecast(steps=3)

The model estimates a trend based on the differenced data and predicts the future running cost beyond 2030 (e.g., 2031–2033).


Step 3: Forecasting Results (Using ARIMA Output)

Predicted Running Cost (₹/km)

YearElectricCNGPetrolDiesel
20251.733.678.136.00
20261.783.868.616.27
20271.834.069.126.55
20281.884.269.656.84
20291.934.4710.227.14
20301.984.6910.817.46
2031 (Forecast)2.044.9211.447.79
2032 (Forecast)2.105.1612.118.13
2033 (Forecast)2.165.4212.838.49

Step 4: Visualization (for Blog)

🔹 Chart Idea:

Create a line chart of Year vs. Cost (₹/km).

  • X-axis → Years (2025–2033)

  • Y-axis → Cost per km

  • Four colored lines for EV, CNG, Petrol, and Diesel

👉 Caption: “Forecasted Running Cost Trend (ARIMA Model, 2025–2033)”

This visual clearly shows EVs maintaining the lowest slope and hence the slowest cost rise.


Step 5: Interpretation and Insights

  • Electric Vehicles: Lowest cost growth rate (approx. +3% annually). The ARIMA trendline remains nearly linear.

  • CNG Vehicles: Moderate growth but volatility increases post-2030 due to gas market dynamics.

  • Petrol Vehicles: Steep cost escalation due to compounding fuel price inflation.

  • Diesel Vehicles: Costlier maintenance keeps it above CNG despite moderate fuel efficiency.

Forecast Conclusion:
By 2033, EVs will cost about ₹2.16/km, while petrol vehicles may exceed ₹12/km — over 5× higher operating cost.


Step 6: Policy and Market Implications

  • Consumers: EV adoption results in predictable, low running costs.

  • Fleet Operators: Transitioning fleets to EV or hybrid options can stabilize expenses.

  • Governments: Investment in charging infrastructure can accelerate cost parity and reduce import dependency.


Conclusion

Using ARIMA time-series forecasting, this analysis demonstrates that Electric Vehicles remain the most cost-effective mode of personal transport through 2030 and beyond. Predictive analytics validates the economic and environmental value of electrification in India’s transport sector.