Introduction to time series analysis (data science)
WHAT :
As we can normally guess, time series data refers to data collected chronologically . What makes it different from others is that it each variable is dependent on previous variables and there are time specific variations which are peculiar and part of normal trend (temporal variations). These are used to make predictions and analysis.
Firstly , one should understand basic terminologies of time — series data.
BASIC TERMS:
- Trend: increase or decrease of data over time. eg. gold prices -upward trend.
- Seasonality: Patterns that repeat at smaller periods. For example, sales of winter clothing .
- Seasonal Adjustment : Process of removing seasonality from time series data.
3. Cyclical Patterns: recurring behavior generally spanning multiple years. eg. election cycles leading to fluctuations in economic indicators.
4. Stationarity: A property of a time series where statistical properties such as mean, variance, and autocorrelation remain constant over time. Stationary time series are easier to model and analyze.
5. Autocorrelation: The correlation between observations at different time points within a time series.
6.A longitudinal time series refers to a set of data points that are collected over time from the same individuals, entities, or subjects over time. eg.individual trajectories of students’ academic performance.
7.Parallel time series refer to a collection of data points taken at the same time intervals but from different individuals, entities, or subjects over time. eg. academic performance levels between different schools at specific time points.
8. Secular trend — long-term trend or the underlying trend, spanning multiple years or decades much more than cyclical. eg. global warming positive trend.
Now we came across basic terms , we’ll see how we make the time series analysis work in the next article.
