What is Time? What is Forecasting?

Time is the river in which we all float. We can't see it, touch it, or stop it, but we feel its passage every second of every day. As young explorers, your first mission is to understand this mysterious river—not with complicated physics, but with the simple, powerful tool of recording. Imagine having a notebook where you write down one thing that happens to you each day: the temperature outside, how much allowance you get, or your highest score in a video game. After a week, you'll have a list of numbers with dates. That list is called a time series, and it's your time machine. By looking at patterns in this list, you can make forecasts—educated guesses about what might happen tomorrow, next week, or even next month. Forecasting isn't magic; it's the art and science of using the past to navigate the future. In this chapter, we'll dive into what time really means, how we chop it into measurable pieces, and how those pieces form the building blocks of prediction. We'll also start your own Time Diary, because the best forecasts begin with your own data.

What Is Time, Really?

Before we can predict the future, we must understand the stage on which all events play out: time itself. What is time? If you ask a philosopher, you might hear that time is an illusion, a human construct to make sense of change. Ask a physicist, and they'll talk about spacetime, relativity, and the arrow of entropy. But for our journey as young forecasters, we don't need to solve these deep mysteries. We'll treat time as a practical tool: something we can measure, record, and use to order events. At its core, time is a way to sequence moments—to say that one thing happens before another. That sequence is crucial because forecasting relies on the idea that the past influences the future. If events were completely random with no order, prediction would be impossible.

Think about your day. You wake up, eat breakfast, go to school, play with friends, do homework, and go to bed. These events happen in a specific order, and we use time to label that order. But time feels different in different situations. Have you ever noticed how waiting for your birthday seems to take forever, but the last day of vacation flies by in a blur? That's because our perception of time is subjective—it speeds up when we're having fun and slows down when we're bored. However, for forecasting, we need an objective, consistent measure. That's why humans invented clocks and calendars.

Clocks break time into small, equal pieces: seconds, minutes, hours. Calendars break it into larger pieces: days, weeks, months, years. These units are based on natural cycles: a day is one rotation of Earth, a year is one orbit around the Sun. But why do we have weeks and months? Weeks come from ancient Babylonians who liked the number 7 (seven celestial bodies they saw). Months are roughly based on the Moon's cycle (about 29.5 days). So our time units are a mix of astronomy and history. For forecasting, the key is that we agree on these units so we can record events consistently. When we say "Monday" or "3:30 PM," everyone knows what we mean.

Now, what does it mean to "record" an event? We attach a time stamp to it. A time stamp is like a label that says exactly when something happened. For example: "2023-10-15 08:00: Temperature 18°C." That tells us the date, the hour, and the temperature. If we record many such stamps, we get a collection of time-stamped data. But for forecasting, we usually want the data in a specific order: from earliest to latest. That ordered list is a time series. A time series is simply a sequence of data points collected at regular intervals over time. The "regular intervals" part is important: if we measure temperature every day at noon, that's regular. If we measure randomly whenever we think of it, that's irregular and harder to analyze.

Let's make this concrete. Suppose you decide to track your daily mood on a scale of 1 to 10 (1 = very sad, 10 = very happy). You record it every evening before bed. After a week, you might have:

• Monday: 7
• Tuesday: 8
• Wednesday: 6
• Thursday: 9
• Friday: 10
• Saturday: 9
• Sunday: 8

That's a time series: mood scores with corresponding dates, ordered chronologically. You can look at this series and see patterns: maybe you're happier on Fridays (weekend anticipation!) and lower on Wednesdays (midweek slump). Those patterns help you predict future moods. If next Wednesday is coming, you might guess a 6 again. That's forecasting in action!

But time series aren't just about moods. They're everywhere:

• **Daily temperature**: The high temperature each day, recorded by weather stations.
• **Weekly allowance**: The amount of money you receive each week, and how much you spend.
• **Monthly video game scores**: Your best score in a game each month as you practice.
• **Hourly traffic**: Number of cars passing a point each hour.
• **Yearly population**: How many people live in your city each year.

All these are sequences of numbers tied to time. And all can be forecasted if we find patterns.

> Sidebar: Time in Different Cultures > Not all cultures think about time the same way! Some indigenous groups see time as circular—seasons repeat, life cycles renew—while Western cultures often see it as a straight line from past to future. This affects how people record and think about events. For forecasting, we use the linear view because it helps us project forward.

Measuring Time: From Sunlight to Smartphones

How do we actually chop time into pieces we can record? The history of timekeeping is a wild adventure. Ancient humans first noticed time through natural cycles: the Sun rising and setting (days), the Moon phases (months), and the changing seasons (years). They built sundials to track the Sun's shadow, water clocks that dripped at a steady rate, and hourglasses with sand. But these were imprecise.

The big leap came with mechanical clocks in the Middle Ages, then pendulum clocks in the 17th century, which were much more accurate. Today, the most accurate clocks are atomic clocks, which use the vibration of atoms (like cesium) to define the second. The international standard, Coordinated Universal Time (UTC), is based on these atomic clocks, with occasional "leap seconds" added to account for Earth's slowing rotation.

For our forecasting journey, we don't need atomic precision. We just need consistent intervals. Common intervals include:

• **Yearly**: Annual revenue of a company, population census every 10 years.
• **Monthly**: Rainfall totals, electricity bills.
• **Weekly**: Sales at a lemonade stand, number of new books read.
• **Daily**: Stock market prices, steps taken, temperature.
• **Hourly**: Website visitors, heart rate during exercise.
• **Minutely**: Sound volume in a room, speed of a car.

The interval you choose depends on what you're measuring and how quickly it changes. If you're tracking your mood, daily makes sense because moods can fluctuate daily. If you're tracking your height, monthly or yearly is better because height changes slowly.

When you record a time series, you typically use a table with two columns: one for the time stamp (date/time) and one for the value (what you're measuring). For example:

| Date | Temperature (°C) | |------------|------------------| | 2023-10-01 | 18 | | 2023-10-02 | 19 | | 2023-10-03 | 20 | | 2023-10-04 | 21 | | 2023-10-05 | 22 | | 2023-10-06 | 23 | | 2023-10-07 | 24 |

This table shows daily high temperatures for a week. Notice the dates are consecutive and the temperature seems to increase by 1°C each day. That pattern is a trend—a long-term increase or decrease. Trends are one of the most important patterns in forecasting.

But what if your data isn't perfectly regular? Sometimes we miss a day, or we record at slightly different times each day. For basic forecasting, we try to keep inte