Skip to main content
SimilarExport ZIP
100%

timeline - seasonality detection (I)

✓ Published0🌍 Public
KKcnarf
Last edited Dec 3, 2018
Created on Feb 11, 2016

This interactive visualization tests a time series for seasonality by comparing it against a shifted version of itself. The left panel displays the original series and its lagged copy as draggable points, while the right panel plots the correlation coefficient for different shift amounts. Users can adjust the suspected season length and the magnitude of the seasonal component. The D3 v3 code computes autocorrelation values, visually linking the lagged timeline, the correlation plot, and a bar chart. Data comes from a CSV file and is rendered as SVG with animated transitions.

AI-generated description

If you suspect a time serie has a seasonality component, this is an example of how to validate/invalidate this hypothesis.

Seasonality means that the time serie has a periodic component, repeating the same pattern on each period. For example, sales of a store may have a week-based seasonality: sales increase on saturday, while there is no sale at all on sunday.

Graphically speaking, detecting a seasonality is (quite) easy: just look for a repeating pattern. Note that it could be difficult if the pattern has a long period, or/and the order of magnitude of the seasonilaty is low (ie. lowest and highest values are not so far from the season's mean, but in this case there might be no seasonality at all ! ).

Computationnaly speaking, one can use the autocorrelation technique. Correlation is a technique that allows to determine if 2 time series are correlated (ie. if they behave in the same way). Autocorrelation is the same technique applied to only 1 time serie: it allows to compare the correlation between a time serie and the same time serie shifted by a certain amount of time. Hence, if you suspect a seasonality of 7 days, then the autocorrelation between the time serie and the same time serie shifted by 7 days will validate (autocorrelation near 1) or invalidate (autocorrelation near 0) the seasonality.

Usages :

  • in the left graph, Drag & Drop points to update the timeline and create seasons of your choice (below the graph are some shortcuts)
  • then test for a particular seasonality period, and see how the coefficient of correlation behaves;
  • decrease the order of magnitude of the season component to see that when this order is small, then it becomes difficult to detect a season; this is because the coefficient of correlation is constantly high; even for the test of a suspected season's length that is not the real one.

Notes:

  • another <a href='http://bl.ocks.org/Kcnarf/1e6da47724c39156adb3'>block</a> explains time series correlation
  • another <a href='http://bl.ocks.org/Kcnarf/0a8fe1caa2ac025c8e86'>block</a> deals with the impact of seasonality when computing the trend of a timeline

Acknowledgments:

  • done with D3 v3.5.5
  • <a href='http://blockbuilder.org'>blockbuilder.org</a>
mit Licensed

Similar vizzes

Loading thumbnail…

timeline - seasonality detection (II)

This interactive D3.v3 visualization demonstrates how to detect seasonality in a time series using a correlogram, showing both the timeline and its autocorrelation coefficients. Users can drag data points to modify the timeline, adjust season length, seasonality magnitude, and trend via controls. The left panel displays the timeline with draggable dots, while the right panel shows correlation bars for lags 1-10. The example reads from a CSV file and uses d3.behavior.drag and d3.svg.axis to enable manipulation and rendering.

KKcnarf
92% match
Loading thumbnail…

timeline - seasonality detection (III)

This interactive D3 v3 visualization demonstrates why detrending a time series before computing a correlogram is essential for detecting seasonality. It shows a synthetic timeline with adjustable season length, seasonality magnitude, and overall trend, alongside a correlation bar chart for each lag. Users can toggle detrending; with the raw series, strong trends mask seasonal patterns, while detrended data reveals even small seasonal cycles clearly. The code uses D3’s linear scales, SVG rendering, drag behavior, and transition animations, with data generated inline and links to adjust parameters.

KKcnarf
88% match
Loading thumbnail…

timeline - trend & seasonality

An interactive D3.v3 timeline visualizes the impact of seasonality on a global trend. Users can drag individual data points, invert, or permute the seasonality to explore how internal seasonal trends affect the overall slope. The chart displays raw values, deseasonalized season means (large dots), and a least-squares trend line, all within an SVG with animation. Dragging points updates the trend in real time, demonstrating how a windowed-mean approach mitigates seasonal effects.

KKcnarf
84% match
Loading thumbnail…

timeline - correlation

This example demonstrates how to assess the correlation between two time series. It shows a line chart for each series and a scatter plot of one series against the other, where the alignment of points indicates the strength of correlation. A dynamic coefficient of correlation is displayed, computed using the least squares method. Users can interactively drag data points, adjust trends, and modify dispersion to see their impact, with d3.svg transitions animating the changes.

KKcnarf
72% match
Loading thumbnail…

timeline - trend

The visualization demonstrates how a single outlier or a user-dragged point dramatically alters a least-squares trend line. It shows a scatter plot of a 20-point time series, where each blue dot can be moved with the mouse to immediately recalculate and redraw the steel-blue regression line. A timeline path connects the points, and buttons allow inverting the trend or creating an outlier. Built with D3.v3, it uses `d3.behavior.drag`, `d3.svg.axis`, and `transition` functions to animate updates, with data loaded from a CSV via `d3.csv`.

KKcnarf
72% match
Loading thumbnail…

timeline - FFT decomposition

This interactive D3.js v4 visualization demonstrates how a timeline can be decomposed into its constituent frequencies using a Fast Fourier Transform (FFT). Users can drag and drop data points in the top graph to modify the time series, or use shortcuts to adjust seasonality period, seasonal magnitude, and trend. The FFT algorithm, implemented from Rosetta Code's JavaScript version, calculates frequency components displayed as bars in a lower panel. Clicking components applies an inverse FFT to reconstruct a filtered timeline, showing which frequencies contribute to the original signal.

KKcnarf
71% match
Loading thumbnail…

timeline - comparing trends

This example shows how to compare trends between two time series by presenting raw and relative (mean-reduced) data side-by-side. The left graph displays raw values, while the right graph shows each value divided by its series mean, allowing users to assess whether absolute or proportional changes are more meaningful. Drag-and-drop interactions enable manipulation of individual data points or entire timelines, with the linked reduced trends updating in real-time. Built with D3 v3.5.5, the visualization uses SVG for rendering, d3.behavior.drag for interactions, and a least-squares regression line computed for each trend.

KKcnarf
70% match
Loading thumbnail…

timeline - trend, confidence interval, outliers

This interactive example demonstrates how a trend line and its 95% confidence interval respond to changes in time-series data. Users can drag individual data points to see their immediate impact on both the trend and the confidence interval, or drag the entire timeline to show that time-axis shifts have no effect. The visualization also provides controls to increase or decrease data dispersion, which alters the confidence interval but not the trend, and to force an outlier. The trend is computed using the least squares method, with the confidence interval based on ±1.96 standard deviations. Built with D3 v3.5.5, the SVG rendering uses `d3.behavior.drag` for interactions and transitions for smooth animations.

KKcnarf
68% match