How to Simplify a Line Chart: A Taxonomy and Feature-Based Evaluation
Organizations: University of Utah
Abstract
Simplifying large or noisy time-series is often a prerequisite for producing interpretable line charts, yet existing methods vary widely in how they alter visual features such as trends, extrema, periodicity, curvature, roughness, and overall complexity. We introduce a feature-based taxonomy of line chart simplification that evaluates 19 commonly used algorithms across three families: Transformers, Reducers, and Aggregators. We also formalize 12 perceptually motivated, computationally defined visual features commonly used when analyzing time-series in line charts. We assess preservation using multiple similarity metrics per feature to capture different aspects of fidelity, yielding 21 distinct feature-preservation measurements. To capture the trade-off between fidelity and simplicity, we introduce a Feature--Complexity (FC) score that combines feature preservation with complexity. Our systematic evaluation on 80 datasets reveals distinct family-level patterns: Transformers such as Gaussian, Mean, and Median filters achieve A-mode grades on nearly all metrics but retain full temporal resolution; Reducers such as Douglas-Peucker and M4 receive predominantly D-F mode grades trading fidelity for a reduction in the number of rendered points; and Aggregators ASAP and PAA occupy a middle ground with primarily B-C mode grades. We further characterize the predictability of algorithm behavior across datasets: some methods produce consistent grades regardless of dataset type (Gaussian, Mean, Douglas-Peucker), while others exhibit high data-dependency, with performance varying substantially based on dataset characteristics (FPCS, ASAP). An interactive website supporting exploration of all evaluated techniques, datasets, and results is available at https://feature-taxonomy.vercel.app/.
Figures & tables
| Method | Description | Examples | Main Parameter | Complexity |
|---|---|---|---|---|
| Transformer | Keeps all time steps but replaces each value with a smoothed or filtered estimate | Gaussian filter Mean filter Median filter Min filter Max filter Savitzky-Golay FFT cutoff filter Butterworth Chebyshev Elliptical | standard deviation window length window length window length window length poly window length cutoff frequency cutoff frequency cutoff frequency cutoff frequency | |
| Reducer | Selects a subset of original points and discards the rest while keeping exact values of retained points | Uniform subsampling Douglas-Peucker (RDP) LTTB MinMaxLTTB M4 FPCS TopoLines | sampling interval target size target size target size target size sampling rate persistence threshold | |
| Aggregator | Reduces and transforms by summarizing groups of points into new representative values | ASAP PAA | target resolution number of bins |
| Domain | Dataset Groups | Description | Length |
|---|---|---|---|
| Astronomy | astro | 5 data files; with frequency measurement of radio waves emitted from extraterrestrial matter [ 4 ] | 1947 |
| Climate | climate_awnd climate_tmax climate_prcp | 18 data files; includes average wind speed (awnd), high temperature (tmax), and total precipitation (prcp) from 6 U.S. airports [ 36 ] | 3651 |
| Crime | chi_homicide_weekly chi_homicide_monthly | 2 data files; weekly and monthly homicide counts in Chicago from January 2001 through July 2019 [ 12 ] | 969 222 |
| EEG | eeg_500 eeg_2500 eeg_10000 | 18 data files; EEG signals from 6 channels per window length (500, 2500, 10000) recorded during a visual attention task [ 16 ] | 500 2500 10000 |
| Flights | flights_daily flights_weekly flights_monthly | 3 data files; daily, weekly, and monthly U.S. flight volumes from January 1, 1988 through December 31 [ 7 ] | 7671 1095 252 |
| New Zealand Tourism | nz_tourist_monthly nz_tourist_annual | 2 data files; monthly and annual tourist counts from April 1921 through April 2018 [ 37 ] | 1165 96 |
Data Visualization
Data visualization studies how visual representations help people explore, interpret, communicate, and reason about data.
See all 126 papers