Why Time Series Databases Are the Future of Your Data

Why a Time Series Database?

If you only keep a single state value, your database will become almost useless. You need a time series database.

Data Is the Gold Mine of the Last Decade

Data has been a gold mine for the past ten years, and its growth accelerates each year thanks to related fields such as machine learning and IoT.

Collecting more data from users is valuable: you can study behavior, discover patterns, and imagine new possibilities. In the future, your data will become even more powerful.

Time‑Series Era

You will be able to analyze the past, the present, and the future! Unlike the old approach that only keeps the latest state, you retain the full history.

You generate more data every minute, thanks to better hardware, storage, and smarter algorithms. Data is the standard for everything.

Time‑Series Data Is Everywhere

Even if you don’t think you own such data, you should consider it from a broader perspective.

S&P 500 net assets over time (time series):

Weather data changes daily, and you may want to keep the entire history:

A simple weather forecast showcase really easily a time series data type everyone uses daily. Photo from Environment Canada (via Alex Hickey).

Having a complete history can yield incredible results, such as tracking a stolen Tesla or even your personal Tesla location as time‑series data.

Predict My Future!

Analysis: discover trends over a period.

Visualization: power dashboards for the whole organization.

Machine Learning: more inputs and outputs enable future ML models.

Automation: define thresholds that trigger pipelines or workflows.

What Does Time‑Series Data Look Like?

What Changes?

In the past you mainly performed UPDATE s. With time series you mainly use INSERT s.

Previously you wrote randomly distributed data. With time series you write data for the most recent time interval.

Earlier you wrote based on a primary key. With time series you write using a primary key that includes a timestamp value.

How to Get Started?

You can dive deep in this field by finding a new idea that keeps historical versions of values in your application database.

First step: try to find a time‑series database available on your preferred cloud provider. Next, populate it with sample data formatted for time series—perhaps from a Kaggle competition dealing with time‑series analysis.

After reading this short introduction to time‑series data, consider the final thought: Is all data time‑series data?

Want to Study Time‑Series Further?

Here are a few books, mainly for machine‑learning developers, that can give you deeper knowledge and ideas:

Practical Time Series Analysis: Forecasting with Statistics and Machine Learning (https://amzn.to/393uDzA)

Introduction to Time Series and Forecasting (https://amzn.to/3pGzKvF)

References and Resources

(This article is translated from Ahmed Sakr’s “Time Series Databases Are the Future for Your Data”. Original link: https://medium.com/better-programming/time-series-databases-are-the-future-for-your-data-664c5edb2cde)