The library was designed to create aesthetic, animated (so far only linear) charts based on a given input. The main assumptions of the library were to create smooth transitions between subsequent data sets. For this reason, we have discovered a shortage of existing libraries related to the charts. The current package was created as part of the Rainbow.me project and for this reason it was not designed as a complete and comprehensive solution for displaying various types of charts. However, we will be now using more charts in the whole application, so we believe that the number of functionalities in the application will gradually grow.

Additionally, we are open to new Pull Requests. We want this library to become popular and complete thanks to community activity.

It's a part of the Rainbow.me project.

1. Install react-native-reanimated in the newest version.

yarn add @rainbow-me/animated-charts

The library has been verified with 2.0.0-alpha.6 version of reanimated.

Using TurboModules might have an impact on your current development flow and most likely you don't want to decrease your DX. Since we're not using reanimated in other places in the app, we made some tweaks to disable charts in development mode with compilation macros on iOS. You can find it here

Also, because we're using libraries which currently do not support reanimated 2, we patched exports in reanimated

Furthermore, we found few differences in how the Animated module works with and without TurboModules support, so we made a trick to fallback to the not-TM version of Animated.

Most likely, you don't need any of those patches.

We made a generic example to show briefly what's possible to achieve with this library. A Real-life example is available inside Rainbow!

In order to run an example clone this repo and navigate to Example then:

yarn && cd ios && pod install && cd ..
react-native run-android
react-native run-ios

The library has been designed to provide as much flexibility as possible with the component-based API for easy integration with existing applications.

import React from 'react';
import {Dimensions, View} from 'react-native';
import {ChartDot, ChartPath, ChartPathProvider, monotoneCubicInterpolation} from '@rainbow-me/animated-charts';

export const {width: SIZE} = Dimensions.get('window');

export const data = [
  {x: 1453075200, y: 1.47}, {x: 1453161600, y: 1.37},
  {x: 1453248000, y: 1.53}, {x: 1453334400, y: 1.54},
  {x: 1453420800, y: 1.52}, {x: 1453507200, y: 2.03},
  {x: 1453593600, y: 2.10}, {x: 1453680000, y: 2.50},
  {x: 1453766400, y: 2.30}, {x: 1453852800, y: 2.42},
  {x: 1453939200, y: 2.55}, {x: 1454025600, y: 2.41},
  {x: 1454112000, y: 2.43}, {x: 1454198400, y: 2.20},
];

const points = monotoneCubicInterpolation({data, range: 40});

const BasicExample = () => (
  <View style={{ backgroundColor: 'black' }}>
    <ChartPathProvider data={{ points, smoothingStrategy: 'bezier' }}>
      <ChartPath height={SIZE / 2} stroke="yellow" width={SIZE} />
      <ChartDot style={{ backgroundColor: 'blue' }} />
    </ChartPathProvider>
  </View>
);

The code above generates the chart below:

The whole chart's structure has to be wrapped with ChartProvider. It's responsible for data managing and itself does not have a visual impact on the layout. Under the hood, it uses context API to simplify manipulation with other components. The rule is to use one data series for each wrapper. It's important to have it as a separated component because under this component it's possible useChartData hook and put ChartDot component. However, if you don't need those features, you can pass data directly to ChartPath and drop this wrapper.

• points is an array containing Points to be displayed. A Point is an object containing x and y as a number.
• originalData is an array of points that will not be drawn. However, if you used some strategy of interpolating data or simplifying you might want to present data slightly different from the real one. Then if you'd like labels to be fully correct you may want to provide real data before adjusting them.
• smoothingStrategy. While presenting points path can be drawn with different approaches.
  • If smoothingStrategy is not provided (or set to any other value but for listed here), connects points using linear interpolation.
  • The bezier strategy connects points with a bezier path inspired by d3 shape. It's not parametrized by smoothingFactor.
  • The complex strategy uses approach explained here using cubic splines. It's parametrized by smoothingFactor.
  • The simple strategy is a bit simplified complex strategy using quadratic splines. It's parametrized by smoothingFactor.
• smoothingFactor. Is a value from 0 to 1 defining how smooth a presentation should be. 0 means no smoothing, and it's the default. smoothingFactor has an impact if smoothingStrategy is simple or complex.
• yRange. 2-elements array with numbers describing minY and maxY in this order. When not applied, charts are automatically adjusted vertically. However, when data are monotonical, they might present incorrectly without bounding due to numerical issues.

This component is used for showing the path itself.

Component for displaying the dot for scrubbing on the chart.

Labels are useful while moving finger through the chart to show the exact value in given point.

CurrentPositionVerticalLine is a vertical line moving with user's finger through chart.

OpeningPositionHorizontalLine is a horizontal static line on height of the first point.

TODO

TODO

It's not essential in the library, but we have decided to include a lot of helpers we are (or we were) using for displaying charts.

We have two interpolators which share the most of the API: bSplineInterpolation and monotoneCubicInterpolation.

import { bSplineInterpolation as interpolator } from '@rainbow-me/animated-charts';
// import { monotoneCubicInterpolation as interpolator } from '@rainbow-me/animated-charts';

const interpolatedData = interpolator({data, range: 80});

Code above generates 80 equidistant points from given dataset. Generator accepts object of parameters as an argument:

• data - input of points.

• range is the number of points of the output.

• includeExtremes. If it's vital to include extremes in the output, set to true. However, the data might not be fully equidistant.

• removePointsSurroundingExtremes. Makes sense only if includeExtremes set to true. When disabled, it might be possible that extremes look very "pointy". To get rid of this, you can remove points surrounding extremes. E.g.

  • removePointsSurroundingExtremes = false

    o---------o----Min--o---------o---------o---------o---------o

  • removePointsSurroundingExtremes = true

    o--------------Min------------o---------o---------o---------o

bSplineInterpolation is inspired by victorian lib and uses B-spline interpolation of a given degree.

This curve is inspired by d3 shape. "Produces a cubic spline that preserves monotonicity in y, assuming monotonicity in x, as proposed by Steffen in A simple method for monotonic interpolation in one dimension: “a smooth curve with continuous first-order derivatives that passes through any given set of data points without spurious oscillations. Local extrema can occur only at grid points where they are given by the data, but not in between two adjacent grid points.”

This helper takes only one point per pickRange. Might be useful for very dense data. If it's important, it's possible to include extremes with the includeExtremes flag. E.g.

pickRange = 3, includeExtremes = true

X are equidistant in this case

Y:0          1          7          2         -3          0          1          2
  S----------o----------E----------X----------E----------o----------X----------o----------S

X - points picked because index%3=0

S – the first and the last points are always included.

E – extremes.

It's a bit advanced hook and contains information about current position of dot and corresponding values. Content of the result is changing often, so it's not worth deep documentation now.

The library has been released in a production-ready version. We use it inside the Rainbow.me project so it's verified for use in production. However, it relies on React Native Reanimated 2.0 in the alpha version thus it might not work perfectly. Test it deeply before using it. Until the stable release of Reanimated 2, I think it's worth not marking this library as stable. Although the library works with Reanimated without any changes, we faced a few issues related to our (quite advanced) usage of the library. Thus we made some hacks we're not very proud of and it's for 99% something you should not do. However, if you see some crashes, you may try one of our hacks.

There're a few things left to make it polished regarding linear charts:

• Support for gestures - pinching, swiping, etc.
• more parameters regarding interpolation, smoothing, and animations (i.e. allow to override springConfig and timingConfig)

Reanimated library is licensed under The MIT License.