The study of time series always boils down to the problem of causality: **how did the past affect the future?** In this three-part series, we’ll cover everything about time series in a very detailed manner.

**You can read the further parts here:**

**The mathematics beneath [Part 2/4]****Autocorrelation, Heteroskedasticity, ARMA, ARIMA and more [Part 3/4]**

Due to the vast development of such data by, for instance, the Internet of Things, the digitalization of healthcare, and the growth of smart cities, time-series data and its interpretation are increasingly significant.

We should expect the number, consistency, and significance of time series data to rise exponentially in the coming years.

We will survey a few historical examples of time series data and analysis in these fields in today’s article:

- Medicine
- Weather
- Economics
- Cosmology

## Understanding Time Series Data in Different Scenario

**Definition.** Analysis of time series is the attempt to derive useful description and statistical details from points ordered in chronological order. It is performed both to diagnose past actions and to forecast possible actions.

Time series research developments are the product of new methods of collecting, documenting, and computer visualisation.

### 1. Time series in medicine

Medicine is a data-driven field that has contributed to human understanding for a few decades with interesting time series research.

The first usage of statistics in medicine is quite late. In 1662, John Grauntâ€™s actuarial tables became one of the first findings of the type of reasoning applied to medical questions in the time series.

These **life** **tables **illustrate the possibility of a person of a given age dying before their next birthday.

The study of time series found its way into medicine as the first functional **electrocardiograms** (ECGs) were invented in 1901 to diagnose cardiac problems by measuring the electrical signals flowing through the heart.

### 2. Time series in weather

The telegraph allowed for rapid compilations of atmospheric conditions in time series from several different places in the late 19th century, hundreds of years since many atmospheric measurements had come into use.

By the **1870s**, this method became commonplace in many parts of the world and led to the development of the first significant data sets based on what was occurring in other geographic areas to forecast local weather.

Many countries are actually making increasingly granular weather observations from hundreds or even thousands of weather stations around the globe, and these forecasts are dependent on data with detailed knowledge about the position and facilities of weather stations.

### 3. Time series in economics

Output and market performance metrics have been offering useful data to analyze from a time series perspective for a long time.

The topic of predicting future economic states dependent on the past was most important and urgent. These predictions aren’t only good for making money – They also help foster growth and avert social catastrophes.

In the late 19th and early 20th century, **economic forecasting** arose out of the anxiety caused by episodic financial crises in the United States and Europe.

At the time, the notion that the economy could be compared to a cyclical pattern, much like the weather was thought to behave, was motivated by pioneers and scholars alike.

With the correct calculations, forecasts could be made and market crashes averted, it was thought.

**Richard Dennis** was one of the pioneers of mechanical trading, or time series forecasting via algorithm. He was a self-made millionaire who, by teaching them some chosen laws on how and when to deal, famously converted average citizens into star traders.

### 4. Time series in astronomy

Over time, astronomy has always focused heavily on plotting objects, directions, and measurements.

As early as 800 BC, the sunspot time series was documented in ancient China, making the compilation of sunspot data one of the most well-recorded natural phenomena ever.

A great pragmatist was George Box, a pioneering statistician who helped to create a common time series model. “**All the models are inaccurate, but some are helpful**” he famously said.

In response to a prevalent mindset that proper time series modelling was a matter of finding the right model to match the results, Box made this comment.

The belief that any concept would represent the real world is very doubtful, as he explained.

## Conclusion

This is just one of a whole tutorial from beginning to end on time series analysis. So check out the second part here:

**Time Series and Machine Learning – The mathematics beneath [Part 2/4] **