agakshay / Time-Series-analysis-and-sales-forecast

Time series analysis on automotive dataset and forecasting the sale using SARIMA

Time-Series-analysis-and-sales-forecast

Time series analysis on automotive dataset and forecasting the sales using SARIMA

1. The data has been read using pandas.
2. The data was cleansed and only important data has been kept.
3. The values are plotted to understand the trend over the years.
4. To understand whether the graph is stationary or not, stationarity testing using the Augmented Dickey-Fuller Unit Root Test has been conducted.
5. Energy Spectral Density (ESD) plot for the new cars sales. ESD measures signal energy distribution across frequency.
6. Statistical analysis to understand the probability distribution of data using histogram and qq plot.
7. Decomposition of the data into Seasonal,trend and noise and visualising each plot for better understanding.
8. Autocorrelation and Partial Autocorrelation are plotted for understanding strength of a relationship with an observation in a time series with observations at prior time steps.
9. Training the the model , SARIMA.
10. Sales forecasting and visualisation.

Libraries used:

pandas

numpy

matplotlib

statsmodels

Dataset: https://open.canada.ca/data/en/dataset/0decc62b-3047-417c-81e7-d8f96fac09a9