## What is time-series data?

According to Tableau(an American interactive data visualization software company focused on business intelligence), time-series data is defined as data that is recorded over consistent intervals of time. Let us understand this using some examples. Weather & temperature keep changing with time. The stock market keeps fluctuating with time. So any data coming from places like this is time series. For a better understanding, let us relate this to the recent Covid scenario. Covid cases in some countries increased exponentially with time and some countries tried to take measures quickly and took control of the situation. This data again is time series only. This data can be used to understand the trend in which covid affected certain countries and the present covid data can be used to understand how covid will increase/decrease in the future and the rate at which it will increase/decrease can be predicted too.

## Research

To understand more about the use case that I mentioned above you can refer to my research paper iieta.org/journals/ria/paper/10.18280/ria.3.. that got published by IIETA(International Information and Engineering Technology Association). My research basically involves the usage of “Prophet” (an open-source software by the Facebook data science team) to analyze & predict the covid cases in different countries and compare these stats of different countries and understand how some countries did well in controlling the covid situation quickly. You can install prophet as a library in python and further use it to deduce inferences from time-series data.

## Time Series Analysis Vs Time Series Prediction

Analysis means examining something by finding information from the details provided to us. Prediction means to foretell the feature beforehand. For one to predict something he/she needs to have some solid proof in the current situation which he/she predicts what is going to happen in the future. This is exactly what happens in Time series analysis and time series prediction.

Before getting to know more about Time series analysis and prediction. Let us understand certain terms.

## Autocorrelation

According to my understanding, It can be defined as the degree of similarity for a variable at different points of time lets say it rains today so it gets cold outside so there is a chance of tomorrow being cold outside tomorrow too. We cannot surely say that it can be cold after 30 days too. So here temperature is a variable and it is similar in days closer to each other rather than temperature after 30 days. Thus, the variable is autocorrelated.

## Trend

It is basically a pattern showing the movement of the series over a long period of time, it can be an increasing trend or a decreasing trend.

## Seasonal Variation

It is defined as the predictable movement(or) change/variation in a time series in one year or less.

That is it for this blog, in the next blog, we’ll try to understand in-depth time series analysis and also understand methods to make time-series predictions.