# How to figure out if the problem is time series Forecasting or not?

Recently I have encountered a time-series based, where I have a dataset which contains data for call-centers performance. The dataset contains information about the number of calls made by customers for each date which is further categorized region wise.

Time        Region  Number of Calls
2018-03-01  X       1245
2018-03-02  Y       1390


The task at hand is to forecast the overall performance of call-centers as well as forecasting each region performance as well for the next 3 months to see which regions will perform better than others. I have a dataset with dates of June-2017 to Current Date

Questions

1. Is the data enough to perform a time-series analysis with credible results for the next quarter?
2. How should I tackle the task to forecast Region Wise Performance as well

## 1 Answer

1 - Is the data enough to perform a time-series analysis with credible results for the next quarter?

This dataset is about 700 lines per Region, correct? Assuming you have 5 regions that is 3500 samples. Well this is surely not a large dataset but you can try to perform forecasting models on it as it is not small either. If the model is relatively simple your model may be able to run it. In DS the only way to know what will work is by trying or by having someone that already tried that saying it work, or it don't.

2 - How should I tackle the task to forecast Region Wise Performance as well

Region Wise, you have even less data, but still you can perform it in the same way, and probably you can get this as a bonus from the general model for all regions.

Your only challenge is to encode the Region into a way that can be used for ARIMA, NNs or other regressive models.

Maybe data from the Region is more relevant to a shallow model than the region itself (such as number of employers, population size, level of technological advance ...)