Time Series Data

Data Visualization
Prepare for Class 11
Time Series Data
Author

Eli Evans

Published

December 2, 2022

Dependency

  • Temporal dependency: future observations are impacted by previous observations

  • Spatial dependency: observed points being further away from each other or closer together than expected from random data

Time Series Concepts

  • ARIMA: autoregressive integrated moving average, model used for predicting future observations based on past observations, assumes future model follows from past so unexpected changes would not be predicted

    • p: # of autoregressive terms

    • d: # of nonseasonal differences needed for stationarity

    • q: # of lagged forecast errors in prediction equation

  • Stationary: observation properties do not depend of time observed

  • First difference: changes from period at time t-1 to time t

  • First order autoregression: using one previous value to predict future values in time-series data

Charts for Plotting Time

  • A line chart can be used to show the continuous change in a y-axis variable over time plotted on the x-axis, with line color or line type denoting relevant categorical variables

  • An animation can be used for any plot type to show change in time with each frame corresponding to the data at one time

  • A line plot can be used with data mapped on an x-y plane where each subsequent data point is observed at a different time and neighbor data points are connected by a line to show continuity