Data Science Stack Exchange is a question and answer site for Data science professionals, Machine Learning specialists, and those interested in learning more about the field. It only takes a minute to sign up.
Sign up to join this community
I am working on building a time series model, but the dataset I have only has date features for the year; the month and date are not available. What would be a suitable model to use and is it even possible? The dataset is an Excel file from different years I have merged into a single sheet with records for each year arranged in alphabetical order.
The first question about applying time-series models is whether you can detect some patterns by yourself or not.
For instance, if you find that the values rise every last week of the month, then you might expect time-series models to be useful.
But if your data is too coarse to extract anything interesting because of external and unpredictable impacts (ex: economic crisis, accident in several stations, etc.) the added value would be too limited.
In addition, you should gather enough data to teach the model patterns. Consequently, if you only have weekly data, you should have at least 3 years (~150 records) to teach the model some patterns due to seasonality and the impact of special events (ex: stock market crash).
You could add some external data like weather or other products associated with your values. This would improve the model prediction quality.
